fdo-mirrors/NetworkManager
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/NetworkManager/NetworkManager.git synced 2025-12-21 00:30:07 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
NetworkManager/clients/cli/connections.c

8924 lines
302 KiB
C
Raw Normal View History

cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
/* nmcli - command-line tool to control NetworkManager
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with this program; if not, write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
* Copyright 2010 - 2015 Red Hat, Inc.
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
*/
all: cleanup includes and let "nm-default.h" include "config.h" - All internal source files (except "examples", which are not internal) should include "config.h" first. As also all internal source files should include "nm-default.h", let "config.h" be included by "nm-default.h" and include "nm-default.h" as first in every source file. We already wanted to include "nm-default.h" before other headers because it might contains some fixes (like "nm-glib.h" compatibility) that is required first. - After including "nm-default.h", we optinally allow for including the corresponding header file for the source file at hand. The idea is to ensure that each header file is self contained. - Don't include "config.h" or "nm-default.h" in any header file (except "nm-sd-adapt.h"). Public headers anyway must not include these headers, and internal headers are never included after "nm-default.h", as of the first previous point. - Include all internal headers with quotes instead of angle brackets. In practice it doesn't matter, because in our public headers we must include other headers with angle brackets. As we use our public headers also to compile our interal source files, effectively the result must be the same. Still do it for consistency. - Except for <config.h> itself. Include it with angle brackets as suggested by https://www.gnu.org/software/autoconf/manual/autoconf.html#Configuration-Headers
2016-02-19 14:57:48 +01:00
#include "nm-default.h"
build: do the right thing with config.h (bgo #644664) 1) it shouldn't be included in headers 2) it should be the first thing included in source files 3) it's needed for getting translation right
2011-03-14 01:00:56 -05:00
cli: move NmcMetaGenericInfo to "utils.h"
2017-04-04 13:52:13 +02:00
#include "connections.h"
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
cli: wait a bit even when '--nowait' is specified; for permissions check (rh #614866) NetworkManager checks clients' permissions on requests, so nmcli mustn't quit too early to give NM a chance to reach it on D-Bus.
2010-07-29 16:16:20 +02:00
#include <unistd.h>
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
#include <errno.h>
#include <signal.h>
#include <netinet/ether.h>
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
#include <readline/readline.h>
#include <readline/history.h>
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: move utils function from common.h to nm-meta-setting-desc.c These functions are only used by nm-meta-setting-desc.c. Make them internal. Unfortunately, they are part of "common.h" which cannot be used without the rest of nmcli. Still todo.
2017-03-28 12:16:31 +02:00
#include "nm-client-utils.h"
cli: move NmcMetaGenericInfo to "utils.h"
2017-04-04 13:52:13 +02:00
#include "nm-vpn-helpers.h"
#include "nm-meta-setting-access.h"
#include "nm-secret-agent-simple.h"
cli: move utils function from common.h to nm-meta-setting-desc.c These functions are only used by nm-meta-setting-desc.c. Make them internal. Unfortunately, they are part of "common.h" which cannot be used without the rest of nmcli. Still todo.
2017-03-28 12:16:31 +02:00
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
#include "utils.h"
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
#include "common.h"
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
#include "settings.h"
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
#include "devices.h"
cli: add 'nmcli agent' command (bgo #739568) Synopsis: nmcli agent { secret | polkit | all } The command runs separate NetworkManager secret agent or session polkit agent, or both. It is useful when - no other secret agent is available (such as GUI nm-applet, gnome-shell, KDE applet) - no other polkit agent is available (such as polkit-gnome-authentication-agent-1, polkit-kde-authentication-agent-1 or lxpolkit) https://bugzilla.gnome.org/show_bug.cgi?id=739568
2014-10-30 15:45:43 +01:00
#include "polkit-agent.h"
cli: use designated initializers for setting NmcOutputField fields Otherwise, changing the structure is difficult because it all depends on the order (and you don't immdiately see which field is used where). Also, drop the name_l10n field.
2017-03-30 11:06:18 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
typedef struct _OptionInfo OptionInfo;
struct _OptionInfo {
const char *setting_name;
const char *property;
const char *option;
enum {
OPTION_NONE = 0x00,
OPTION_REQD = 0x01, /* Don't ask to ask. */
OPTION_DONT_ASK = 0x02, /* Don't ask interactively by default */
OPTION_MULTI = 0x04, /* Ask multiple times, do an append instead of set. */
OPTION_DISABLED = 0x10, /* Don't ask due to runtime decision. */
OPTION_ENABLED = 0x20, /* Override OPTION_DONT_ASK due to runtime decision. */
} flags;
const char *prompt;
const char *def_hint;
gboolean (*check_and_set)(NmCli *nmc, NMConnection *connection, OptionInfo *option, const char *value, GError **error);
rl_compentry_func_t *generator_func;
};
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* define some prompts for connection editor */
#define EDITOR_PROMPT_SETTING _("Setting name? ")
#define EDITOR_PROMPT_PROPERTY _("Property name? ")
#define EDITOR_PROMPT_CON_TYPE _("Enter connection type: ")
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
/* define some other prompts */
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
#define PROMPT_CON_TYPE N_("Connection type")
cli: tab-complete "Interface name [*]" in questionnaire mode
2016-09-15 10:32:44 +02:00
#define PROMPT_IFNAME N_("Interface name [*]")
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
#define PROMPT_VPN_TYPE N_("VPN type")
#define PROMPT_MASTER N_("Master")
#define PROMPT_IB_MODE N_("Transport mode")
#define WORD_DATAGRAM "datagram"
#define WORD_CONNECTED "connected"
#define PROMPT_IB_MODE_CHOICES "(" WORD_DATAGRAM "/" WORD_CONNECTED ") [" WORD_DATAGRAM "]"
#define PROMPT_BT_TYPE N_("Bluetooth type")
#define WORD_PANU "panu"
#define WORD_DUN_GSM "dun-gsm"
#define WORD_DUN_CDMA "dun-cdma"
#define PROMPT_BT_TYPE_CHOICES "(" WORD_PANU "/" WORD_DUN_GSM "/" WORD_DUN_CDMA ") [" WORD_PANU "]"
#define PROMPT_BOND_MODE N_("Bonding mode")
#define PROMPT_BOND_MON_MODE N_("Bonding monitoring mode")
#define WORD_MIIMON "miimon"
#define WORD_ARP "arp"
#define PROMPT_BOND_MON_MODE_CHOICES "(" WORD_MIIMON "/" WORD_ARP ") [" WORD_MIIMON "]"
#define PROMPT_ADSL_PROTO N_("Protocol")
#define PROMPT_ADSL_PROTO_CHOICES "(" NM_SETTING_ADSL_PROTOCOL_PPPOA "/" NM_SETTING_ADSL_PROTOCOL_PPPOE "/" NM_SETTING_ADSL_PROTOCOL_IPOATM ")"
#define PROMPT_WIFI_MODE N_("Wi-Fi mode")
#define WORD_INFRA "infrastructure"
#define WORD_AP "ap"
#define WORD_ADHOC "adhoc"
#define PROMPT_WIFI_MODE_CHOICES "(" WORD_INFRA "/" WORD_AP "/" WORD_ADHOC ") [" WORD_INFRA "]"
#define PROMPT_ADSL_ENCAP N_("ADSL encapsulation")
#define PROMPT_ADSL_ENCAP_CHOICES "(" NM_SETTING_ADSL_ENCAPSULATION_VCMUX "/" NM_SETTING_ADSL_ENCAPSULATION_LLC ") [none]"
#define PROMPT_TUN_MODE N_("Tun mode")
#define WORD_TUN "tun"
#define WORD_TAP "tap"
#define PROMPT_TUN_MODE_CHOICES "(" WORD_TUN "/" WORD_TAP ") [" WORD_TUN "]"
#define PROMPT_IP_TUNNEL_MODE N_("IP Tunnel mode")
#define PROMPT_MACVLAN_MODE N_("MACVLAN mode")
cli: fix some TAB completions for nmcli -a con add (cherry picked from commit 1cee85f556f7771e93c8637dac5ad284a405bf20)
2017-01-21 21:12:49 +01:00
#define PROMPT_MACSEC_MODE N_("MACsec mode")
#define WORD_PSK "psk"
#define WORD_EAP "eap"
#define PROMPT_MACSEC_MODE_CHOICES "(" WORD_PSK "/" WORD_EAP ")"
#define PROMPT_PROXY_METHOD N_("Proxy method")
#define WORD_NONE "none"
#define WORD_AUTO "auto"
#define PROMPT_PROXY_METHOD_CHOICES "(" WORD_NONE "/" WORD_AUTO ") [" WORD_NONE "]"
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
#define PROMPT_CONNECTION _("Connection (name, UUID, or path)")
#define PROMPT_VPN_CONNECTION _("VPN connection (name, UUID, or path)")
#define PROMPT_CONNECTIONS _("Connection(s) (name, UUID, or path)")
#define PROMPT_ACTIVE_CONNECTIONS _("Connection(s) (name, UUID, path or apath)")
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
build: avoid compiler warnings about non-constant format strings (-Wformat-nonliteral)
2016-06-05 11:46:06 +02:00
#define BASE_PROMPT "nmcli> "
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
/*****************************************************************************/
const NmcMetaGenericInfo *const nmc_fields_con_show[] = {
NMC_META_GENERIC ("NAME"), /* 0 */
NMC_META_GENERIC ("UUID"), /* 1 */
NMC_META_GENERIC ("TYPE"), /* 2 */
NMC_META_GENERIC ("TIMESTAMP"), /* 3 */
NMC_META_GENERIC ("TIMESTAMP-REAL"), /* 4 */
NMC_META_GENERIC ("AUTOCONNECT"), /* 5 */
NMC_META_GENERIC ("AUTOCONNECT-PRIORITY"), /* 6 */
NMC_META_GENERIC ("READONLY"), /* 7 */
NMC_META_GENERIC ("DBUS-PATH"), /* 8 */
NMC_META_GENERIC ("ACTIVE"), /* 9 */
NMC_META_GENERIC ("DEVICE"), /* 10 */
NMC_META_GENERIC ("STATE"), /* 11 */
NMC_META_GENERIC ("ACTIVE-PATH"), /* 12 */
NMC_META_GENERIC ("SLAVE"), /* 13 */
NULL,
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
};
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
#define NMC_FIELDS_CON_SHOW_COMMON "NAME,UUID,TYPE,DEVICE"
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
const NmcMetaGenericInfo *const nmc_fields_con_active_details_general[] = {
NMC_META_GENERIC ("GROUP"), /* 0 */
NMC_META_GENERIC ("NAME"), /* 1 */
NMC_META_GENERIC ("UUID"), /* 2 */
NMC_META_GENERIC ("DEVICES"), /* 3 */
NMC_META_GENERIC ("STATE"), /* 4 */
NMC_META_GENERIC ("DEFAULT"), /* 5 */
NMC_META_GENERIC ("DEFAULT6"), /* 6 */
NMC_META_GENERIC ("SPEC-OBJECT"), /* 7 */
NMC_META_GENERIC ("VPN"), /* 8 */
NMC_META_GENERIC ("DBUS-PATH"), /* 9 */
NMC_META_GENERIC ("CON-PATH"), /* 10 */
NMC_META_GENERIC ("ZONE"), /* 11 */
NMC_META_GENERIC ("MASTER-PATH"), /* 12 */
NULL,
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
};
cli: add property-info for the rest
2017-03-26 14:30:17 +02:00
#define NMC_FIELDS_SETTINGS_NAMES_ALL NM_SETTING_CONNECTION_SETTING_NAME","\
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
NM_SETTING_WIRED_SETTING_NAME","\
NM_SETTING_802_1X_SETTING_NAME","\
NM_SETTING_WIRELESS_SETTING_NAME","\
NM_SETTING_WIRELESS_SECURITY_SETTING_NAME","\
NM_SETTING_IP4_CONFIG_SETTING_NAME","\
NM_SETTING_IP6_CONFIG_SETTING_NAME","\
NM_SETTING_SERIAL_SETTING_NAME","\
NM_SETTING_PPP_SETTING_NAME","\
NM_SETTING_PPPOE_SETTING_NAME","\
adsl: add libnm-glib and nmcli code For the finale, this is the libnm-glib and nmcli part of the support, with this you can now make a full pppoatm connection from NetworkManager by configuring it in system scope in a keyfile. As an example, here is mine: (password and username snipped for obvious reasons) [connection] id=MYISP uuid=34d04e69-fdd9-4231-af2c-25ed1f34dc1e type=adsl timestamp=1304621332 [adsl] username=user@myisp.com password=mypassword vpi=8 vci=35 encapsulation=vcmux protocol=pppoa [ipv6] method=ignore [ipv4] method=auto Extract from logs: NetworkManager[29155]: <info> Activation (ueagle-atm0) starting connection 'MYISP' NetworkManager[29155]: <info> (ueagle-atm0): device state change: disconnected -> prepare (reason 'none') [30 40 0] NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 1 of 5 (Device Prepare) scheduled... NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 1 of 5 (Device Prepare) started... NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 2 of 5 (Device Configure) scheduled... NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 1 of 5 (Device Prepare) complete. NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 2 of 5 (Device Configure) starting... NetworkManager[29155]: <info> (ueagle-atm0): device state change: prepare -> config (reason 'none') [40 50 0] NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 2 of 5 (Device Configure) successful. NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 3 of 5 (IP Configure Start) scheduled. NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 2 of 5 (Device Configure) complete. NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 3 of 5 (IP Configure Start) started... NetworkManager[29155]: <info> (ueagle-atm0): device state change: config -> ip-config (reason 'none') [50 70 0] NetworkManager[29155]: <info> starting PPP connection NetworkManager[29155]: <debug> [1304671146.590156] [nm-ppp-manager.c:1047] nm_ppp_manager_start(): command line: /usr/sbin/pppd nodetach lock nodefaultroute user user@myisp.com plugin pppoatm.so 8.35 noipdefault noauth usepeerdns lcp-echo-failure 0 lcp-echo-interval 0 ipparam /org/freedesktop/NetworkManager/PPP/0 plugin /opt/nmadsl/lib/pppd/2.4.5/nm-pppd-plugin.so NetworkManager[29155]: <info> pppd started with pid 29175 NetworkManager[29155]: <debug> [1304671146.591235] [NetworkManagerUtils.c:816] nm_utils_get_proc_sys_net_value(): (ueagle-atm0): error reading /proc/sys/net/ipv6/conf/ueagle-atm0/accept_ra: (4) Failed to open file '/proc/sys/net/ipv6/conf/ueagle-atm0/accept_ra': No such file or directory NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 3 of 5 (IP Configure Start) complete. Plugin pppoatm.so loaded. Plugin /opt/nmadsl/lib/pppd/2.4.5/nm-pppd-plugin.so loaded. ** Message: nm-ppp-plugin: (plugin_init): initializing ** Message: nm-ppp-plugin: (nm_phasechange): status 3 / phase 'serial connection' NetworkManager[29155]: <debug> [1304671146.612937] [nm-udev-manager.c:621] handle_uevent(): UDEV event: action 'add' subsys 'net' device 'ppp0' NetworkManager[29155]: <debug> [1304671146.613134] [nm-udev-manager.c:525] net_add(): ignoring interface with type 512 NetworkManager[29155]: <debug> [1304671146.617308] [nm-netlink-monitor.c:117] link_msg_handler(): netlink link message: iface idx 35 flags 0x1090 Using interface ppp0 Connect: ppp0 <--> 8.35 ** Message: nm-ppp-plugin: (nm_phasechange): status 5 / phase 'establish' ** Message: nm-ppp-plugin: (nm_phasechange): status 6 / phase 'authenticate' ** Message: nm-ppp-plugin: (get_credentials): passwd-hook, requesting credentials... NetworkManager[29155]: <debug> [1304671149.639511] [nm-agent-manager.c:1044] nm_agent_manager_get_secrets(): Secrets requested for connection /org/freedesktop/NetworkManager/Settings/0 (adsl) NetworkManager[29155]: <debug> [1304671149.639684] [nm-settings-connection.c:717] nm_settings_connection_get_secrets(): (34d04e69-fdd9-4231-af2c-25ed1f34dc1e/adsl:1) secrets requested flags 0x1 hint 'password' NetworkManager[29155]: <debug> [1304671149.640950] [nm-agent-manager.c:959] get_start(): (0x9b4ad10/adsl) system settings secrets sufficient NetworkManager[29155]: <debug> [1304671149.641332] [nm-settings-connection.c:573] agent_secrets_done_cb(): (34d04e69-fdd9-4231-af2c-25ed1f34dc1e/adsl:1) existing secrets returned NetworkManager[29155]: <debug> [1304671149.641541] [nm-settings-connection.c:579] agent_secrets_done_cb(): (34d04e69-fdd9-4231-af2c-25ed1f34dc1e/adsl:1) secrets request completed NetworkManager[29155]: <debug> [1304671149.643074] [nm-settings-connection.c:618] agent_secrets_done_cb(): (34d04e69-fdd9-4231-af2c-25ed1f34dc1e/adsl:1) new agent secrets processed ** Message: nm-ppp-plugin: (get_credentials): got credentials from NetworkManager PAP authentication succeeded ** Message: nm-ppp-plugin: (nm_phasechange): status 8 / phase 'network' local IP address 94.70.90.201 remote IP address 80.106.108.12 primary DNS address 195.170.0.1 secondary DNS address 195.170.2.2 ** Message: nm-ppp-plugin: (nm_phasechange): status 9 / phase 'running' ** Message: nm-ppp-plugin: (nm_ip_up): ip-up event ** Message: nm-ppp-plugin: (nm_ip_up): sending Ip4Config to NetworkManager... NetworkManager[29155]: <debug> [1304671150.607440] [nm-netlink-monitor.c:117] link_msg_handler(): netlink link message: iface idx 35 flags 0x110D1 NetworkManager[29155]: <info> PPP manager(IP Config Get) reply received. NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 4 of 5 (IP4 Configure Get) scheduled... NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 4 of 5 (IP4 Configure Get) started... NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 5 of 5 (IP Configure Commit) scheduled... NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 4 of 5 (IP4 Configure Get) complete. NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 5 of 5 (IP Configure Commit) started... NetworkManager[29155]: <debug> [1304671150.608918] [nm-system.c:222] sync_addresses(): (ppp0): syncing addresses (family 2) NetworkManager[29155]: <info> (ueagle-atm0): device state change: ip-config -> activated (reason 'none') [70 100 0] NetworkManager[29155]: <info> Policy set 'MYISP' (ppp0) as default for IPv4 routing and DNS. NetworkManager[29155]: <info> Activation (ueagle-atm0) successful, device activated. NetworkManager[29155]: <info> Activation (ueagle-atm0) Stage 5 of 5 (IP Configure Commit) complete. Signed-off-by: Pantelis Koukousoulas <pktoss@gmail.com>
2011-05-17 21:03:20 +03:00
NM_SETTING_ADSL_SETTING_NAME","\
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
NM_SETTING_GSM_SETTING_NAME","\
NM_SETTING_CDMA_SETTING_NAME","\
NM_SETTING_BLUETOOTH_SETTING_NAME","\
NM_SETTING_OLPC_MESH_SETTING_NAME","\
cli: add 'infiniband' setting support
2011-12-09 13:20:36 +01:00
NM_SETTING_VPN_SETTING_NAME","\
cli: add 'bond' setting
2012-03-05 07:32:21 +01:00
NM_SETTING_INFINIBAND_SETTING_NAME","\
cli: add 'vlan' setting
2012-03-05 07:50:28 +01:00
NM_SETTING_BOND_SETTING_NAME","\
cli: add bridge support
2012-10-30 17:11:28 -05:00
NM_SETTING_VLAN_SETTING_NAME","\
NM_SETTING_BRIDGE_SETTING_NAME","\
cli: add team driver support Signed-off-by: Jiri Pirko <jiri@resnulli.us>
2013-07-25 15:36:49 +02:00
NM_SETTING_BRIDGE_PORT_SETTING_NAME","\
NM_SETTING_TEAM_SETTING_NAME","\
cli: add support for DCB settings
2013-11-06 08:16:40 -06:00
NM_SETTING_TEAM_PORT_SETTING_NAME"," \
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
NM_SETTING_DCB_SETTING_NAME"," \
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
NM_SETTING_TUN_SETTING_NAME"," \
cli: add macvlan support
2015-09-18 11:03:17 +02:00
NM_SETTING_IP_TUNNEL_SETTING_NAME"," \
cli: macsec support
2016-07-01 11:56:33 +02:00
NM_SETTING_MACSEC_SETTING_NAME"," \
cli: add vxlan support
2015-10-13 18:53:10 +02:00
NM_SETTING_MACVLAN_SETTING_NAME"," \
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
NM_SETTING_VXLAN_SETTING_NAME"," \
NM_SETTING_PROXY_SETTING_NAME
cli: add property-info for the rest
2017-03-26 14:30:17 +02:00
// NM_SETTING_DUMMY_SETTING_NAME
// NM_SETTING_WIMAX_SETTING_NAME
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
const NmcMetaGenericInfo *const nmc_fields_con_active_details_vpn[] = {
NMC_META_GENERIC ("GROUP"), /* 0 */
NMC_META_GENERIC ("TYPE"), /* 1 */
NMC_META_GENERIC ("USERNAME"), /* 2 */
NMC_META_GENERIC ("GATEWAY"), /* 3 */
NMC_META_GENERIC ("BANNER"), /* 4 */
NMC_META_GENERIC ("VPN-STATE"), /* 5 */
NMC_META_GENERIC ("CFG"), /* 6 */
NULL,
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
};
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
const NmcMetaGenericInfo *const nmc_fields_con_active_details_groups[] = {
NMC_META_GENERIC_WITH_NESTED ("GENERAL", nmc_fields_con_active_details_general + 1), /* 0 */
NMC_META_GENERIC_WITH_NESTED ("IP4", nmc_fields_ip4_config + 1), /* 1 */
NMC_META_GENERIC_WITH_NESTED ("DHCP4", nmc_fields_dhcp4_config + 1), /* 2 */
NMC_META_GENERIC_WITH_NESTED ("IP6", nmc_fields_ip6_config + 1), /* 3 */
NMC_META_GENERIC_WITH_NESTED ("DHCP6", nmc_fields_dhcp6_config + 1), /* 4 */
NMC_META_GENERIC_WITH_NESTED ("VPN", nmc_fields_con_active_details_vpn + 1), /* 5 */
NULL,
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
};
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
/* Pseudo group names for 'connection show <con>' */
/* e.g.: nmcli -f profile con show my-eth0 */
/* e.g.: nmcli -f active con show my-eth0 */
#define CON_SHOW_DETAIL_GROUP_PROFILE "profile"
#define CON_SHOW_DETAIL_GROUP_ACTIVE "active"
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: remove useless static function prototypes
2011-09-06 15:30:15 +02:00
/* glib main loop variable - defined in nmcli.c */
extern GMainLoop *loop;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
static guint progress_id = 0; /* ID of event source for displaying progress */
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
/* for readline TAB completion in editor */
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
typedef struct {
NmCli *nmc;
cli: fix TAB-completion and 'goto' after saving connection On 'save' we replace settings in the local connection, so we also need to update stored setting pointers.
2014-01-20 17:38:11 +01:00
char *con_type;
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
NMConnection *connection;
NMSetting *setting;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
const char *property;
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
} TabCompletionInfo;
static TabCompletionInfo nmc_tab_completion = {NULL, NULL, NULL, NULL};
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
static char *gen_connection_types (const char *text, int state);
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
static void
usage (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection { COMMAND | help }\n\n"
nmcli: add missing commands to 'nmcli connection help' usage description
2016-07-14 23:39:16 +02:00
"COMMAND := { show | up | down | add | modify | clone | edit | delete | monitor | reload | load | import | export }\n\n"
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
" show [--active] [--order <order spec>]\n"
cli: add global --show-secrets option - it controls echoing passwords input on terminal - it replaces --show-secrets in 'nmcli connection show', which is deprecated now - it replaces --show-password in 'nmcli device wifi hotspot', which is deprecated now
2015-12-04 13:29:42 +01:00
" show [--active] [id | uuid | path | apath] <ID> ...\n\n"
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
" up [[id | uuid | path] <ID>] [ifname <ifname>] [ap <BSSID>] [passwd-file <file with passwords>]\n\n"
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
" down [id | uuid | path | apath] <ID> ...\n\n"
cli: add master option to "nmcli c add"
2015-06-08 15:18:06 +02:00
" add COMMON_OPTIONS TYPE_SPECIFIC_OPTIONS SLAVE_OPTIONS IP_OPTIONS [-- ([+|-]<setting>.<property> <value>)+]\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" modify [--temporary] [id | uuid | path] <ID> ([+|-]<setting>.<property> <value>)+\n\n"
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
" clone [--temporary] [id | uuid | path ] <ID> <new name>\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" edit [id | uuid | path] <ID>\n"
" edit [type <new_con_type>] [con-name <new_con_name>]\n\n"
" delete [id | uuid | path] <ID>\n\n"
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
" monitor [id | uuid | path] <ID> ...\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" reload\n\n"
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
" load <filename> [ <filename>... ]\n\n"
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
" import [--temporary] type <type> file <file to import>\n\n"
" export [id | uuid | path] <ID> [<output file>]\n\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
static void
usage_connection_show (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection show { ARGUMENTS | help }\n"
"\n"
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
"ARGUMENTS := [--active] [--order <order spec>]\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"\n"
"List in-memory and on-disk connection profiles, some of which may also be\n"
"active if a device is using that connection profile. Without a parameter, all\n"
"profiles are listed. When --active option is specified, only the active\n"
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
"profiles are shown. --order allows custom connection ordering (see manual page).\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"\n"
cli: add global --show-secrets option - it controls echoing passwords input on terminal - it replaces --show-secrets in 'nmcli connection show', which is deprecated now - it replaces --show-password in 'nmcli device wifi hotspot', which is deprecated now
2015-12-04 13:29:42 +01:00
"ARGUMENTS := [--active] [id | uuid | path | apath] <ID> ...\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"\n"
"Show details for specified connections. By default, both static configuration\n"
"and active connection data are displayed. It is possible to filter the output\n"
"using global '--fields' option. Refer to the manual page for more information.\n"
"When --active option is specified, only the active profiles are taken into\n"
cli: add global --show-secrets option - it controls echoing passwords input on terminal - it replaces --show-secrets in 'nmcli connection show', which is deprecated now - it replaces --show-password in 'nmcli device wifi hotspot', which is deprecated now
2015-12-04 13:29:42 +01:00
"account. Use global --show-secrets option to reveal associated secrets as well.\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
static void
usage_connection_up (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection up { ARGUMENTS | help }\n"
"\n"
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
"ARGUMENTS := [id | uuid | path] <ID> [ifname <ifname>] [ap <BSSID>] [nsp <name>] [passwd-file <file with passwords>]\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"\n"
"Activate a connection on a device. The profile to activate is identified by its\n"
"name, UUID or D-Bus path.\n"
"\n"
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
"ARGUMENTS := ifname <ifname> [ap <BSSID>] [nsp <name>] [passwd-file <file with passwords>]\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"\n"
"Activate a device with a connection. The connection profile is selected\n"
"automatically by NetworkManager.\n"
"\n"
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
"ifname - specifies the device to active the connection on\n"
"ap - specifies AP to connect to (only valid for Wi-Fi)\n"
"nsp - specifies NSP to connect to (only valid for WiMAX)\n"
"passwd-file - file with password(s) required to activate the connection\n\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
static void
usage_connection_down (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection down { ARGUMENTS | help }\n"
"\n"
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
"ARGUMENTS := [id | uuid | path | apath] <ID> ...\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"\n"
"Deactivate a connection from a device (without preventing the device from\n"
"further auto-activation). The profile to deactivate is identified by its name,\n"
"UUID or D-Bus path.\n\n"));
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
static void
usage_connection_add (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection add { ARGUMENTS | help }\n"
"\n"
cli: add master option to "nmcli c add"
2015-06-08 15:18:06 +02:00
"ARGUMENTS := COMMON_OPTIONS TYPE_SPECIFIC_OPTIONS SLAVE_OPTIONS IP_OPTIONS [-- ([+|-]<setting>.<property> <value>)+]\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" COMMON_OPTIONS:\n"
" type <type>\n"
" ifname <interface name> | \"*\"\n"
" [con-name <connection name>]\n"
cli: fix formating of "nmcli con add help" output
2015-07-15 16:08:31 +02:00
" [autoconnect yes|no]\n"
" [save yes|no]\n"
cli: add master option to "nmcli c add"
2015-06-08 15:18:06 +02:00
" [master <master (ifname, or connection UUID or name)>]\n"
" [slave-type <master connection type>]\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" TYPE_SPECIFIC_OPTIONS:\n"
" ethernet: [mac <MAC address>]\n"
" [cloned-mac <cloned MAC address>]\n"
" [mtu <MTU>]\n\n"
" wifi: ssid <SSID>\n"
" [mac <MAC address>]\n"
" [cloned-mac <cloned MAC address>]\n"
cli: Add mode option for wifi connecitons Only default (infrastructure) mode connections can be created and as it's not possible to write mode=ap connections with ifcfg-rh plugins, they can't be switched to mode=ap.
2014-10-04 17:07:42 +02:00
" [mtu <MTU>]\n"
" [mode infrastructure|ap|adhoc]\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" wimax: [mac <MAC address>]\n"
" [nsp <NSP>]\n\n"
" pppoe: username <PPPoE username>\n"
" [password <PPPoE password>]\n"
" [service <PPPoE service name>]\n"
" [mtu <MTU>]\n"
" [mac <MAC address>]\n\n"
" gsm: apn <APN>\n"
" [user <username>]\n"
" [password <password>]\n\n"
" cdma: [user <username>]\n"
" [password <password>]\n\n"
" infiniband: [mac <MAC address>]\n"
" [mtu <MTU>]\n"
" [transport-mode datagram | connected]\n"
" [parent <ifname>]\n"
" [p-key <IPoIB P_Key>]\n\n"
" bluetooth: [addr <bluetooth address>]\n"
" [bt-type panu|dun-gsm|dun-cdma]\n\n"
cli: remove stray space from help description
2015-12-16 13:35:23 +01:00
" vlan: dev <parent device (connection UUID, ifname, or MAC)>\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" id <VLAN ID>\n"
" [flags <VLAN flags>]\n"
" [ingress <ingress priority mapping>]\n"
" [egress <egress priority mapping>]\n"
" [mtu <MTU>]\n\n"
" bond: [mode balance-rr (0) | active-backup (1) | balance-xor (2) | broadcast (3) |\n"
" 802.3ad (4) | balance-tlb (5) | balance-alb (6)]\n"
" [primary <ifname>]\n"
" [miimon <num>]\n"
" [downdelay <num>]\n"
" [updelay <num>]\n"
" [arp-interval <num>]\n"
cli: support new bond option 'lacp_rate' Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-25 12:59:49 +02:00
" [arp-ip-target <num>]\n"
" [lacp-rate slow (0) | fast (1)]\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" bond-slave: master <master (ifname, or connection UUID or name)>\n\n"
" team: [config <file>|<raw JSON data>]\n\n"
" team-slave: master <master (ifname, or connection UUID or name)>\n"
" [config <file>|<raw JSON data>]\n\n"
" bridge: [stp yes|no]\n"
" [priority <num>]\n"
" [forward-delay <2-30>]\n"
" [hello-time <1-10>]\n"
" [max-age <6-40>]\n"
" [ageing-time <0-1000000>]\n"
nmcli: add support for bridge multicast-snooping property
2015-02-20 12:58:30 +01:00
" [multicast-snooping yes|no]\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" [mac <MAC address>]\n\n"
" bridge-slave: master <master (ifname, or connection UUID or name)>\n"
" [priority <0-63>]\n"
" [path-cost <1-65535>]\n"
" [hairpin yes|no]\n\n"
" vpn: vpn-type vpnc|openvpn|pptp|openconnect|openswan|libreswan|ssh|l2tp|iodine|...\n"
" [user <username>]\n\n"
" olpc-mesh: ssid <SSID>\n"
" [channel <1-13>]\n"
" [dhcp-anycast <MAC address>]\n\n"
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
" adsl: username <username>\n"
" protocol pppoa|pppoe|ipoatm\n"
" [password <password>]\n"
" [encapsulation vcmux|llc]\n\n"
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
" tun: mode tun|tap\n"
" [owner <UID>]\n"
" [group <GID>]\n"
" [pi yes|no]\n"
" [vnet-hdr yes|no]\n"
" [multi-queue yes|no]\n\n"
cli: use '|' instead of ',' for delimiting the modes to be consistent with other options
2015-12-17 16:20:10 +01:00
" ip-tunnel: mode ipip|gre|sit|isatap|vti|ip6ip6|ipip6|ip6gre|vti6\n"
cli: document ip-tunnel connection type
2015-12-16 13:52:20 +01:00
" remote <remote endpoint IP>\n"
" [local <local endpoint IP>]\n"
" [dev <parent device (ifname or connection UUID)>]\n\n"
cli: macsec support
2016-07-01 11:56:33 +02:00
" macsec: dev <parent device (connection UUID, ifname, or MAC)>\n"
" mode <psk|eap>\n"
" [cak <key> ckn <key>]\n"
" [encrypt yes|no]\n"
" [port 1-65534]\n\n\n"
cli: remove stray space from help description
2015-12-16 13:35:23 +01:00
" macvlan: dev <parent device (connection UUID, ifname, or MAC)>\n"
cli: add macvlan support
2015-09-18 11:03:17 +02:00
" mode vepa|bridge|private|passthru|source\n"
" [tap yes|no]\n\n"
cli: add vxlan support
2015-10-13 18:53:10 +02:00
" vxlan: id <VXLAN ID>\n"
" remote <IP of multicast group or remote address>\n"
" [local <source IP>]\n"
cli: document ip-tunnel connection type
2015-12-16 13:52:20 +01:00
" [dev <parent device (ifname or connection UUID)>]\n"
cli: add vxlan support
2015-10-13 18:53:10 +02:00
" [source-port-min <0-65535>]\n"
" [source-port-max <0-65535>]\n"
" [destination-port <0-65535>]\n\n"
cli: support dummy connections
2017-02-01 15:19:29 +01:00
" dummy: \n\n"
cli: add master option to "nmcli c add"
2015-06-08 15:18:06 +02:00
" SLAVE_OPTIONS:\n"
" bridge: [priority <0-63>]\n"
" [path-cost <1-65535>]\n"
" [hairpin yes|no]\n\n"
" team: [config <file>|<raw JSON data>]\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" IP_OPTIONS:\n"
" [ip4 <IPv4 address>] [gw4 <IPv4 gateway>]\n"
" [ip6 <IPv6 address>] [gw6 <IPv6 gateway>]\n\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
static void
usage_connection_modify (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection modify { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := [id | uuid | path] <ID> ([+|-]<setting>.<property> <value>)+\n"
"\n"
"Modify one or more properties of the connection profile.\n"
"The profile is identified by its name, UUID or D-Bus path. For multi-valued\n"
"properties you can use optional '+' or '-' prefix to the property name.\n"
"The '+' sign allows appending items instead of overwriting the whole value.\n"
"The '-' sign allows removing selected items instead of the whole value.\n"
"\n"
"Examples:\n"
"nmcli con mod home-wifi wifi.ssid rakosnicek\n"
"nmcli con mod em1-1 ipv4.method manual ipv4.addr \"192.168.1.2/24, 10.10.1.5/8\"\n"
"nmcli con mod em1-1 +ipv4.dns 8.8.4.4\n"
"nmcli con mod em1-1 -ipv4.dns 1\n"
"nmcli con mod em1-1 -ipv6.addr \"abbe::cafe/56\"\n"
"nmcli con mod bond0 +bond.options mii=500\n"
"nmcli con mod bond0 -bond.options downdelay\n\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
static void
usage_connection_clone (void)
{
g_printerr (_("Usage: nmcli connection clone { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := [--temporary] [id | uuid | path] <ID> <new name>\n"
"\n"
"Clone an existing connection profile. The newly created connection will be\n"
"the exact copy of the <ID>, except the uuid property (will be generated) and\n"
"id (provided as <new name> argument).\n\n"));
}
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
static void
usage_connection_edit (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection edit { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := [id | uuid | path] <ID>\n"
"\n"
"Edit an existing connection profile in an interactive editor.\n"
"The profile is identified by its name, UUID or D-Bus path\n"
"\n"
"ARGUMENTS := [type <new connection type>] [con-name <new connection name>]\n"
"\n"
"Add a new connection profile in an interactive editor.\n\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
static void
usage_connection_delete (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection delete { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := [id | uuid | path] <ID>\n"
"\n"
"Delete a connection profile.\n"
"The profile is identified by its name, UUID or D-Bus path.\n\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
static void
usage_connection_monitor (void)
{
g_printerr (_("Usage: nmcli connection monitor { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := [id | uuid | path] <ID> ...\n"
"\n"
"Monitor connection profile activity.\n"
"This command prints a line whenever the specified connection changes.\n"
"Monitors all connection profiles in case none is specified.\n\n"));
}
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
static void
usage_connection_reload (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection reload { help }\n"
"\n"
"Reload all connection files from disk.\n\n"));
cli: add 'help' for second level commands (bgo #698237) (rh #1034119) In addition to 'nmcli connection help', you can now use: nmcli con up help nmcli con down help nmcli dev wifi help etc. The help option can also be in the form of --help/-h, or even shortened. https://bugzilla.gnome.org/show_bug.cgi?id=698237 https://bugzilla.redhat.com/show_bug.cgi?id=1034119
2013-12-18 14:49:06 +01:00
}
static void
usage_connection_load (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Usage: nmcli connection load { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := <filename> [<filename>...]\n"
"\n"
"Load/reload one or more connection files from disk. Use this after manually\n"
"editing a connection file to ensure that NetworkManager is aware of its latest\n"
"state.\n\n"));
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
static void
usage_connection_import (void)
{
g_printerr (_("Usage: nmcli connection import { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := [--temporary] type <type> file <file to import>\n"
"\n"
"Import an external/foreign configuration as a NetworkManager connection profile.\n"
"The type of the input file is specified by type option.\n"
"Only VPN configurations are supported at the moment. The configuration\n"
"is imported by NetworkManager VPN plugins.\n\n"));
}
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
static void
usage_connection_export (void)
{
g_printerr (_("Usage: nmcli connection export { ARGUMENTS | help }\n"
"\n"
"ARGUMENTS := [id | uuid | path] <ID> [<output file>]\n"
"\n"
"Export a connection. Only VPN connections are supported at the moment.\n"
"The data are directed to standard output or to a file if a name is given.\n\n"));
}
cli: split out new file "nm-meta-setting-access.h" "nm-meta-setting-desc.h" contains static type description, vtable and (internal) accessor functions. Add accessor functions that operate on top of the type description to "nm-meta-setting-access.h".
2017-03-31 14:45:46 +02:00
static NMSetting *
nmc_setting_new_for_name (const char *name)
{
GType stype;
NMSetting *setting = NULL;
if (name) {
stype = nm_setting_lookup_type (name);
if (stype != G_TYPE_INVALID) {
setting = g_object_new (stype, NULL);
g_warn_if_fail (NM_IS_SETTING (setting));
}
}
return setting;
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
/* quit main loop */
static void
quit (void)
{
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
if (progress_id) {
g_source_remove (progress_id);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
progress_id = 0;
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
nmc_terminal_erase_line ();
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
g_main_loop_quit (loop); /* quit main loop */
}
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
static char *
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
construct_header_name (const char *base, const char *spec)
{
if (spec == NULL)
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
return g_strdup (base);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
return g_strdup_printf ("%s (%s)", base, spec);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
}
static const char *
active_connection_state_to_string (NMActiveConnectionState state)
{
switch (state) {
case NM_ACTIVE_CONNECTION_STATE_ACTIVATING:
return _("activating");
case NM_ACTIVE_CONNECTION_STATE_ACTIVATED:
return _("activated");
case NM_ACTIVE_CONNECTION_STATE_DEACTIVATING:
return _("deactivating");
case NM_ACTIVE_CONNECTION_STATE_DEACTIVATED:
return _("deactivated");
case NM_ACTIVE_CONNECTION_STATE_UNKNOWN:
default:
return _("unknown");
}
}
static const char *
libnm, core, cli, tui: fix the capitalization of various types GLib/Gtk have mostly settled on the convention that two-letter acronyms in type names remain all-caps (eg, "IO"), but longer acronyms become initial-caps-only (eg, "Tcp"). NM was inconsistent, with most long acronyms using initial caps only (Adsl, Cdma, Dcb, Gsm, Olpc, Vlan), but others using all caps (DHCP, PPP, PPPOE, VPN). Fix libnm and src/ to use initial-caps only for all three-or-more-letter-long acronyms (and update nmcli and nmtui for the libnm changes).
2014-06-26 13:44:36 -04:00
vpn_connection_state_to_string (NMVpnConnectionState state)
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
{
switch (state) {
case NM_VPN_CONNECTION_STATE_PREPARE:
return _("VPN connecting (prepare)");
case NM_VPN_CONNECTION_STATE_NEED_AUTH:
return _("VPN connecting (need authentication)");
case NM_VPN_CONNECTION_STATE_CONNECT:
return _("VPN connecting");
case NM_VPN_CONNECTION_STATE_IP_CONFIG_GET:
return _("VPN connecting (getting IP configuration)");
case NM_VPN_CONNECTION_STATE_ACTIVATED:
return _("VPN connected");
case NM_VPN_CONNECTION_STATE_FAILED:
return _("VPN connection failed");
case NM_VPN_CONNECTION_STATE_DISCONNECTED:
return _("VPN disconnected");
default:
return _("unknown");
}
}
/* Caller has to free the returned string */
static char *
get_ac_device_string (NMActiveConnection *active)
{
GString *dev_str;
const GPtrArray *devices;
int i;
if (!active)
return NULL;
/* Get devices of the active connection */
dev_str = g_string_new (NULL);
devices = nm_active_connection_get_devices (active);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (i = 0; i < devices->len; i++) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
NMDevice *device = g_ptr_array_index (devices, i);
const char *dev_iface = nm_device_get_iface (device);
if (dev_iface) {
g_string_append (dev_str, dev_iface);
g_string_append_c (dev_str, ',');
}
}
if (dev_str->len > 0)
g_string_truncate (dev_str, dev_str->len - 1); /* Cut off last ',' */
return g_string_free (dev_str, FALSE);
}
static NMActiveConnection *
get_ac_for_connection (const GPtrArray *active_cons, NMConnection *connection)
{
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
const char *con_path, *ac_con_path;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
int i;
NMActiveConnection *ac = NULL;
/* Is the connection active? */
con_path = nm_connection_get_path (connection);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (i = 0; i < active_cons->len; i++) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
NMActiveConnection *candidate = g_ptr_array_index (active_cons, i);
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
NMRemoteConnection *con;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
con = nm_active_connection_get_connection (candidate);
cli: fix errors and crash for invisible connections * two users are logged in: user_a and user_b * user_b creates a connection visible only to him 'user_b_private' (permissions: user:user_b) * user_b activates the connection user_b_private * user_a does not see the connection profile, but he does see the active connection * user_a calls nmcli con nmcli con show user_b_private
2014-10-16 16:08:24 +02:00
ac_con_path = con ? nm_connection_get_path (NM_CONNECTION (con)) : NULL;
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
if (!g_strcmp0 (ac_con_path, con_path)) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
ac = candidate;
break;
}
}
return ac;
}
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
/* Put secrets into local connection. */
cli: add '--show-secrets' option for 'nmcli connection show' It can be used to display connection secrets (passwords). When used, it will get secrets for the connection profile and merge it into the connection's settings before displaying it. Example: nmcli con show -s hotel-wifi
2014-09-11 16:34:50 +02:00
static void
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
update_secrets_in_connection (NMRemoteConnection *remote, NMConnection *local)
cli: add '--show-secrets' option for 'nmcli connection show' It can be used to display connection secrets (passwords). When used, it will get secrets for the connection profile and merge it into the connection's settings before displaying it. Example: nmcli con show -s hotel-wifi
2014-09-11 16:34:50 +02:00
{
GVariant *secrets;
int i;
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
GError *error = NULL;
cli: add '--show-secrets' option for 'nmcli connection show' It can be used to display connection secrets (passwords). When used, it will get secrets for the connection profile and merge it into the connection's settings before displaying it. Example: nmcli con show -s hotel-wifi
2014-09-11 16:34:50 +02:00
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
for (i = 0; i < _NM_META_SETTING_TYPE_NUM; i++) {
secrets = nm_remote_connection_get_secrets (remote, nm_meta_setting_infos[i].setting_name, NULL, NULL);
cli: add '--show-secrets' option for 'nmcli connection show' It can be used to display connection secrets (passwords). When used, it will get secrets for the connection profile and merge it into the connection's settings before displaying it. Example: nmcli con show -s hotel-wifi
2014-09-11 16:34:50 +02:00
if (secrets) {
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
if (!nm_connection_update_secrets (local, NULL, secrets, &error) && error) {
g_printerr (_("Error updating secrets for %s: %s\n"),
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
nm_meta_setting_infos[i].setting_name,
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
error->message);
g_clear_error (&error);
}
cli: add '--show-secrets' option for 'nmcli connection show' It can be used to display connection secrets (passwords). When used, it will get secrets for the connection profile and merge it into the connection's settings before displaying it. Example: nmcli con show -s hotel-wifi
2014-09-11 16:34:50 +02:00
g_variant_unref (secrets);
}
}
}
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
static gboolean
cli: display '<hidden>' string for secret properties unless the user explicitly say to show them.
2014-09-29 12:22:39 +02:00
nmc_connection_profile_details (NMConnection *connection, NmCli *nmc, gboolean secrets)
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
{
GError *error = NULL;
GArray *print_settings_array;
cli: allow '--field group.field' for nmcli con show conf <bla> That means you can display single property. Example: $ nmcli -f connection.id,802-3-ethernet.mtu s c my-eth-profile connection.id: my-eth-profile 802-3-ethernet.mtu: auto
2013-12-10 16:23:31 +01:00
GPtrArray *prop_array = NULL;
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
int i;
char *fields_str;
char *fields_all = NMC_FIELDS_SETTINGS_NAMES_ALL;
char *fields_common = NMC_FIELDS_SETTINGS_NAMES_ALL;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
const char *base_hdr = _("Connection profile details");
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
gboolean was_output = FALSE;
if (!nmc->required_fields || strcasecmp (nmc->required_fields, "common") == 0)
fields_str = fields_common;
else if (!nmc->required_fields || strcasecmp (nmc->required_fields, "all") == 0)
fields_str = fields_all;
else
fields_str = nmc->required_fields;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
print_settings_array = parse_output_fields (fields_str, (const NMMetaAbstractInfo *const*) nm_meta_setting_infos_editor_p (),
TRUE, &prop_array, &error);
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
if (error) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
g_string_printf (nmc->return_text, _("Error: 'connection show': %s"), error->message);
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
g_error_free (error);
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
return FALSE;
}
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
g_assert (print_settings_array);
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
/* Main header */
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
{
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
gs_free char *header_name = NULL;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
gs_free NmcOutputField *row = NULL;
gs_unref_array GArray *out_indices = NULL;
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
header_name = construct_header_name (base_hdr, nm_connection_get_id (connection));
out_indices = parse_output_fields (NMC_FIELDS_SETTINGS_NAMES_ALL,
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
(const NMMetaAbstractInfo *const*) nm_meta_setting_infos_editor_p (),
FALSE, NULL, NULL);
row = g_new0 (NmcOutputField, _NM_META_SETTING_TYPE_NUM + 1);
for (i = 0; i < _NM_META_SETTING_TYPE_NUM; i++)
row[i].info = (const NMMetaAbstractInfo *) &nm_meta_setting_infos_editor[i];
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
cli: merge NmcPrintFields into NmcOutputData and pass it directly to print_required_fields()
2017-03-31 12:55:43 +02:00
print_required_fields (&nmc->nmc_config, NMC_OF_FLAG_MAIN_HEADER_ONLY,
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
out_indices, header_name,
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
0, row);
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
}
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
/* Loop through the required settings and print them. */
for (i = 0; i < print_settings_array->len; i++) {
cli: use setting_details() to simplify printing connection details
2013-01-22 14:41:29 +01:00
NMSetting *setting;
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
int section_idx = g_array_index (print_settings_array, int, i);
cli: allow '--field group.field' for nmcli con show conf <bla> That means you can display single property. Example: $ nmcli -f connection.id,802-3-ethernet.mtu s c my-eth-profile connection.id: my-eth-profile 802-3-ethernet.mtu: auto
2013-12-10 16:23:31 +01:00
const char *prop_name = (const char *) g_ptr_array_index (prop_array, i);
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output != NMC_PRINT_TERSE && !nmc->nmc_config.multiline_output && was_output)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("\n"); /* Empty line */
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
was_output = FALSE;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
setting = nm_connection_get_setting_by_name (connection, nm_meta_setting_infos_editor[section_idx].general->setting_name);
cli: use setting_details() to simplify printing connection details
2013-01-22 14:41:29 +01:00
if (setting) {
cli: add nmc_output_selection_create() to parse field selection nmc_output_selection_create() returns a less opaque result then a GArray and a GPtrArray for the groups.
2017-04-03 14:04:28 +02:00
setting_details (&nmc->nmc_config, setting, prop_name, secrets);
cli: use setting_details() to simplify printing connection details
2013-01-22 14:41:29 +01:00
was_output = TRUE;
cli: add bridge support
2012-10-30 17:11:28 -05:00
}
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
}
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
g_array_free (print_settings_array, TRUE);
cli: allow '--field group.field' for nmcli con show conf <bla> That means you can display single property. Example: $ nmcli -f connection.id,802-3-ethernet.mtu s c my-eth-profile connection.id: my-eth-profile 802-3-ethernet.mtu: auto
2013-12-10 16:23:31 +01:00
if (prop_array)
g_ptr_array_free (prop_array, TRUE);
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
cli: make id|uuid specifiers optional for 'connection list' and 'connection status' And also allow: - identifying connetions with 'path'/'apath' accepting either the whole D-Dus path or just an index - multiple connections on cmdline at once nmcli connection list [[id|uuid|path] <ID>] nmcli connection status [[id|uuid|path|apath] <ID>] Examples: nmcli c l nmcli c l "My Wi-Fi 1" airpoirt1 cpath 23 id "My Wi-Fi 2" nmcli c s nmcli c s apath 1 ee19b893-e2a7-42cf-ac82-34506f07a05f
2012-11-16 12:43:44 +01:00
return TRUE;
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
}
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
static NMActiveConnection *
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
find_active_connection (const GPtrArray *active_cons,
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
const GPtrArray *cons,
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
const char *filter_type,
const char *filter_val,
cli/connections: add active connection completion Start completing by the id if the filter type is not specified
2016-07-12 00:23:16 +02:00
int *idx,
gboolean complete)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
int i;
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
int start = (idx && *idx > 0) ? *idx : 0;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
const char *path, *a_path, *path_num, *a_path_num;
const char *id;
const char *uuid;
cli: fix errors and crash for invisible connections * two users are logged in: user_a and user_b * user_b creates a connection visible only to him 'user_b_private' (permissions: user:user_b) * user_b activates the connection user_b_private * user_a does not see the connection profile, but he does see the active connection * user_a calls nmcli con nmcli con show user_b_private
2014-10-16 16:08:24 +02:00
NMRemoteConnection *con;
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
NMActiveConnection *found = NULL;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (i = start; i < active_cons->len; i++) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
NMActiveConnection *candidate = g_ptr_array_index (active_cons, i);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: fix errors and crash for invisible connections * two users are logged in: user_a and user_b * user_b creates a connection visible only to him 'user_b_private' (permissions: user:user_b) * user_b activates the connection user_b_private * user_a does not see the connection profile, but he does see the active connection * user_a calls nmcli con nmcli con show user_b_private
2014-10-16 16:08:24 +02:00
con = nm_active_connection_get_connection (candidate);
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
cli: fix errors and crash for invisible connections * two users are logged in: user_a and user_b * user_b creates a connection visible only to him 'user_b_private' (permissions: user:user_b) * user_b activates the connection user_b_private * user_a does not see the connection profile, but he does see the active connection * user_a calls nmcli con nmcli con show user_b_private
2014-10-16 16:08:24 +02:00
id = nm_active_connection_get_id (candidate);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
uuid = nm_active_connection_get_uuid (candidate);
cli: fix errors and crash for invisible connections * two users are logged in: user_a and user_b * user_b creates a connection visible only to him 'user_b_private' (permissions: user:user_b) * user_b activates the connection user_b_private * user_a does not see the connection profile, but he does see the active connection * user_a calls nmcli con nmcli con show user_b_private
2014-10-16 16:08:24 +02:00
path = con ? nm_connection_get_path (NM_CONNECTION (con)) : NULL;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
path_num = path ? strrchr (path, '/') + 1 : NULL;
cli: fix errors and crash for invisible connections * two users are logged in: user_a and user_b * user_b creates a connection visible only to him 'user_b_private' (permissions: user:user_b) * user_b activates the connection user_b_private * user_a does not see the connection profile, but he does see the active connection * user_a calls nmcli con nmcli con show user_b_private
2014-10-16 16:08:24 +02:00
a_path = nm_object_get_path (NM_OBJECT (candidate));
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
a_path_num = a_path ? strrchr (a_path, '/') + 1 : NULL;
cli: use own code to output connection details (settings) nmcli uses its own code to dump connections' details instead of libnm-util's nm_connection_dump(). It allows to structure output and present values of the settings according to nmcli needs.
2010-03-31 17:14:35 +02:00
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
/* When filter_type is NULL, compare connection ID (filter_val)
* against all types. Otherwise, only compare against the specific
* type. If 'path' or 'apath' filter types are specified, comparison
* against numeric index (in addition to the whole path) is allowed.
*/
cli/connections: add active connection completion Start completing by the id if the filter type is not specified
2016-07-12 00:23:16 +02:00
if (!filter_type || strcmp (filter_type, "id") == 0) {
if (complete)
nmc_complete_strings (filter_val, id, NULL);
if (strcmp (filter_val, id) == 0)
goto found;
}
if (!filter_type || strcmp (filter_type, "uuid") == 0) {
if (complete && (filter_type || *filter_val))
nmc_complete_strings (filter_val, uuid, NULL);
if (strcmp (filter_val, uuid) == 0)
goto found;
}
if (!filter_type || strcmp (filter_type, "path") == 0) {
if (complete && (filter_type || *filter_val))
nmc_complete_strings (filter_val, path, filter_type ? path_num : NULL, NULL);
if (g_strcmp0 (filter_val, path) == 0 || (filter_type && g_strcmp0 (filter_val, path_num) == 0))
goto found;
}
if (!filter_type || strcmp (filter_type, "apath") == 0) {
if (complete && (filter_type || *filter_val))
nmc_complete_strings (filter_val, a_path, filter_type ? a_path_num : NULL, NULL);
if (g_strcmp0 (filter_val, a_path) == 0 || (filter_type && g_strcmp0 (filter_val, a_path_num) == 0))
goto found;
}
continue;
found:
if (!idx)
return candidate;
if (found) {
*idx = i;
return found;
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
}
cli/connections: add active connection completion Start completing by the id if the filter type is not specified
2016-07-12 00:23:16 +02:00
found = candidate;
cli: add 'status' and 'default6' fields to active connections list
2011-12-19 10:09:05 +01:00
}
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
if (idx)
*idx = 0;
return found;
cli: add 'status' and 'default6' fields to active connections list
2011-12-19 10:09:05 +01:00
}
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
void
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
nmc_active_connection_state_to_color (NMActiveConnectionState state, NMMetaTermColor *color)
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
{
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
*color = NM_META_TERM_COLOR_NORMAL;
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
if (state == NM_ACTIVE_CONNECTION_STATE_ACTIVATING)
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
*color = NM_META_TERM_COLOR_YELLOW;
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
else if (state == NM_ACTIVE_CONNECTION_STATE_ACTIVATED)
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
*color = NM_META_TERM_COLOR_GREEN;
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
else if (state > NM_ACTIVE_CONNECTION_STATE_ACTIVATED)
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
*color = NM_META_TERM_COLOR_RED;
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
}
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
static void
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
fill_output_connection (NMConnection *connection, NMClient *client, GPtrArray *output_data, gboolean active_only)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
NMSettingConnection *s_con;
guint64 timestamp;
time_t timestamp_real;
char *timestamp_str;
char *timestamp_real_str = "";
cli: support connection.autoconnect-priority property Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-26 16:30:03 +02:00
char *prio_str;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
NmcOutputField *arr;
NMActiveConnection *ac = NULL;
const char *ac_path = NULL;
const char *ac_state = NULL;
trivial: suppress compiler uninitialized usage warning Suppress a bogus warning on older compilers (gcc 4.7.2)
2015-02-24 19:01:09 -06:00
NMActiveConnectionState ac_state_int = NM_ACTIVE_CONNECTION_STATE_UNKNOWN;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
char *ac_dev = NULL;
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
NMMetaTermColor color;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
s_con = nm_connection_get_setting_connection (connection);
g_assert (s_con);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
ac = get_ac_for_connection (nm_client_get_active_connections (client), connection);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
if (active_only && !ac)
return;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
if (ac) {
ac_path = nm_object_get_path (NM_OBJECT (ac));
nmcli: show connection list in colors Colorize active connections as follows: activated - green activating - yellow deactivating/deactivated - red invisible active connections - dimmed
2015-02-16 10:33:29 +01:00
ac_state_int = nm_active_connection_get_state (ac);
ac_state = active_connection_state_to_string (ac_state_int);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
ac_dev = get_ac_device_string (ac);
}
/* Obtain field values */
timestamp = nm_setting_connection_get_timestamp (s_con);
timestamp_str = g_strdup_printf ("%" G_GUINT64_FORMAT, timestamp);
if (timestamp) {
timestamp_real = timestamp;
timestamp_real_str = g_malloc0 (64);
strftime (timestamp_real_str, 64, "%c", localtime (&timestamp_real));
}
cli: support connection.autoconnect-priority property Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-26 16:30:03 +02:00
prio_str = g_strdup_printf ("%u", nm_setting_connection_get_autoconnect_priority (s_con));
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
arr = nmc_dup_fields_array ((const NMMetaAbstractInfo *const*) nmc_fields_con_show, 0);
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
nmcli: show connection list in colors Colorize active connections as follows: activated - green activating - yellow deactivating/deactivated - red invisible active connections - dimmed
2015-02-16 10:33:29 +01:00
/* Show active connections in color */
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
nmc_active_connection_state_to_color (ac_state_int, &color);
set_val_color_all (arr, color);
nmcli: show connection list in colors Colorize active connections as follows: activated - green activating - yellow deactivating/deactivated - red invisible active connections - dimmed
2015-02-16 10:33:29 +01:00
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
set_val_strc (arr, 0, nm_setting_connection_get_id (s_con));
set_val_strc (arr, 1, nm_setting_connection_get_uuid (s_con));
set_val_strc (arr, 2, nm_setting_connection_get_connection_type (s_con));
set_val_str (arr, 3, timestamp_str);
set_val_str (arr, 4, timestamp ? timestamp_real_str : g_strdup (_("never")));
set_val_strc (arr, 5, nm_setting_connection_get_autoconnect (s_con) ? _("yes") : _("no"));
cli: support connection.autoconnect-priority property Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-26 16:30:03 +02:00
set_val_str (arr, 6, prio_str);
set_val_strc (arr, 7, nm_setting_connection_get_read_only (s_con) ? _("yes") : _("no"));
set_val_strc (arr, 8, nm_connection_get_path (connection));
set_val_strc (arr, 9, ac ? _("yes") : _("no"));
set_val_str (arr, 10, ac_dev);
set_val_strc (arr, 11, ac_state);
set_val_strc (arr, 12, ac_path);
cli: add SLAVE field for 'nmcli con show' output It helps listing or searching for slave connections. For example, one can do $ nmcli -f name,uuid,type,slave connection show | awk '$NF == "bridge"' to show all bridge slave connections.
2016-11-25 23:07:09 +01:00
set_val_strc (arr, 13, nm_setting_connection_get_slave_type (s_con));
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
g_ptr_array_add (output_data, arr);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
}
cli: show active connection even if the connection profile is not visible $ nmcli con show
2014-10-16 17:51:07 +02:00
static void
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
fill_output_connection_for_invisible (NMActiveConnection *ac, GPtrArray *output_data)
cli: show active connection even if the connection profile is not visible $ nmcli con show
2014-10-16 17:51:07 +02:00
{
NmcOutputField *arr;
const char *ac_path = NULL;
const char *ac_state = NULL;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
char *name, *ac_dev = NULL;
cli: show active connection even if the connection profile is not visible $ nmcli con show
2014-10-16 17:51:07 +02:00
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
name = g_strdup_printf ("<invisible> %s", nm_active_connection_get_id (ac));
cli: show active connection even if the connection profile is not visible $ nmcli con show
2014-10-16 17:51:07 +02:00
ac_path = nm_object_get_path (NM_OBJECT (ac));
ac_state = active_connection_state_to_string (nm_active_connection_get_state (ac));
ac_dev = get_ac_device_string (ac);
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
arr = nmc_dup_fields_array ((const NMMetaAbstractInfo *const*) nmc_fields_con_show, 0);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
set_val_str (arr, 0, name);
cli: show active connection even if the connection profile is not visible $ nmcli con show
2014-10-16 17:51:07 +02:00
set_val_strc (arr, 1, nm_active_connection_get_uuid (ac));
set_val_strc (arr, 2, nm_active_connection_get_connection_type (ac));
set_val_strc (arr, 3, NULL);
set_val_strc (arr, 4, NULL);
set_val_strc (arr, 5, NULL);
set_val_strc (arr, 6, NULL);
set_val_strc (arr, 7, NULL);
set_val_strc (arr, 8, NULL);
set_val_strc (arr, 9, _("yes"));
set_val_str (arr, 10, ac_dev);
set_val_strc (arr, 11, ac_state);
set_val_strc (arr, 12, ac_path);
cli: add SLAVE field for 'nmcli con show' output It helps listing or searching for slave connections. For example, one can do $ nmcli -f name,uuid,type,slave connection show | awk '$NF == "bridge"' to show all bridge slave connections.
2016-11-25 23:07:09 +01:00
set_val_strc (arr, 13, NULL);
cli: show active connection even if the connection profile is not visible $ nmcli con show
2014-10-16 17:51:07 +02:00
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
set_val_color_fmt_all (arr, NM_META_TERM_FORMAT_DIM);
nmcli: show connection list in colors Colorize active connections as follows: activated - green activating - yellow deactivating/deactivated - red invisible active connections - dimmed
2015-02-16 10:33:29 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
g_ptr_array_add (output_data, arr);
cli: show active connection even if the connection profile is not visible $ nmcli con show
2014-10-16 17:51:07 +02:00
}
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
static void
fill_output_active_connection (NMActiveConnection *active,
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
GPtrArray *output_data,
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
gboolean with_group,
guint32 o_flags)
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
{
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
NMRemoteConnection *con;
cli: add SLAVE field for 'nmcli con show' output It helps listing or searching for slave connections. For example, one can do $ nmcli -f name,uuid,type,slave connection show | awk '$NF == "bridge"' to show all bridge slave connections.
2016-11-25 23:07:09 +01:00
NMSettingConnection *s_con = NULL;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
const GPtrArray *devices;
GString *dev_str;
cli: add 'status' and 'default6' fields to active connections list
2011-12-19 10:09:05 +01:00
NMActiveConnectionState state;
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
NMDevice *master;
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
const char *con_path = NULL, *con_zone = NULL;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
int i;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
const NMMetaAbstractInfo *const*tmpl;
NmcOutputField *arr;
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
int idx_start = with_group ? 0 : 1;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
con = nm_active_connection_get_connection (active);
if (con) {
con_path = nm_connection_get_path (NM_CONNECTION (con));
s_con = nm_connection_get_setting_connection (NM_CONNECTION (con));
g_assert (s_con != NULL);
con_zone = nm_setting_connection_get_zone (s_con);
}
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
cli: add 'status' and 'default6' fields to active connections list
2011-12-19 10:09:05 +01:00
state = nm_active_connection_get_state (active);
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
master = nm_active_connection_get_master (active);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
/* Get devices of the active connection */
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
dev_str = g_string_new (NULL);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
devices = nm_active_connection_get_devices (active);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (i = 0; i < devices->len; i++) {
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
NMDevice *device = g_ptr_array_index (devices, i);
cli: prevent possible crash if removing devices while nmcli running (bgo #693158)
2013-02-05 10:00:58 +01:00
const char *dev_iface = nm_device_get_iface (device);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: prevent possible crash if removing devices while nmcli running (bgo #693158)
2013-02-05 10:00:58 +01:00
if (dev_iface) {
g_string_append (dev_str, dev_iface);
g_string_append_c (dev_str, ',');
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
if (dev_str->len > 0)
g_string_truncate (dev_str, dev_str->len - 1); /* Cut off last ',' */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
tmpl = (const NMMetaAbstractInfo *const*) nmc_fields_con_active_details_general;
if (!with_group)
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
tmpl++;
cli: fill active connections' data even if we can't get NMConnection This can happen e.g. when D-Bus limits number of replies and there's a lot of connections. Then nm_remote_settings_list_connections() won't return all connections and the connection that was activated can be among the missing connections.
2012-05-22 16:54:44 +02:00
/* Fill field values */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
arr = nmc_dup_fields_array (tmpl, o_flags);
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
if (with_group)
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
set_val_strc (arr, 0, nmc_fields_con_active_details_groups[0]->name);
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
set_val_strc (arr, 1-idx_start, nm_active_connection_get_id (active));
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
set_val_strc (arr, 2-idx_start, nm_active_connection_get_uuid (active));
set_val_str (arr, 3-idx_start, dev_str->str);
set_val_strc (arr, 4-idx_start, active_connection_state_to_string (state));
set_val_strc (arr, 5-idx_start, nm_active_connection_get_default (active) ? _("yes") : _("no"));
set_val_strc (arr, 6-idx_start, nm_active_connection_get_default6 (active) ? _("yes") : _("no"));
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
set_val_strc (arr, 7-idx_start, nm_active_connection_get_specific_object_path (active));
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
set_val_strc (arr, 8-idx_start, NM_IS_VPN_CONNECTION (active) ? _("yes") : _("no"));
set_val_strc (arr, 9-idx_start, nm_object_get_path (NM_OBJECT (active)));
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
set_val_strc (arr, 10-idx_start, con_path);
set_val_strc (arr, 11-idx_start, con_zone);
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
set_val_strc (arr, 12-idx_start, master ? nm_object_get_path (NM_OBJECT (master)) : NULL);
cli: add SLAVE field for 'nmcli con show' output It helps listing or searching for slave connections. For example, one can do $ nmcli -f name,uuid,type,slave connection show | awk '$NF == "bridge"' to show all bridge slave connections.
2016-11-25 23:07:09 +01:00
set_val_strc (arr, 13-idx_start, s_con ? nm_setting_connection_get_slave_type (s_con) : NULL);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
g_ptr_array_add (output_data, arr);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
g_string_free (dev_str, FALSE);
}
typedef struct {
char **array;
guint32 idx;
} FillVPNDataInfo;
static void
fill_vpn_data_item (const char *key, const char *value, gpointer user_data)
{
FillVPNDataInfo *info = (FillVPNDataInfo *) user_data;
info->array[info->idx++] = g_strdup_printf ("%s = %s", key, value);
}
// FIXME: The same or similar code for VPN info appears also in nm-applet (applet-dialogs.c),
// and in gnome-control-center as well. It could probably be shared somehow.
static char *
get_vpn_connection_type (NMConnection *connection)
{
const char *type, *p;
/* The service type is in form of "org.freedesktop.NetworkManager.vpnc".
* Extract end part after last dot, e.g. "vpnc"
*/
type = nm_setting_vpn_get_service_type (nm_connection_get_setting_vpn (connection));
p = strrchr (type, '.');
return g_strdup (p ? p + 1 : type);
}
/* VPN parameters can be found at:
* http://git.gnome.org/browse/network-manager-openvpn/tree/src/nm-openvpn-service.h
* http://git.gnome.org/browse/network-manager-vpnc/tree/src/nm-vpnc-service.h
* http://git.gnome.org/browse/network-manager-pptp/tree/src/nm-pptp-service.h
* http://git.gnome.org/browse/network-manager-openconnect/tree/src/nm-openconnect-service.h
* http://git.gnome.org/browse/network-manager-openswan/tree/src/nm-openswan-service.h
* See also 'properties' directory in these plugins.
*/
static const gchar *
find_vpn_gateway_key (const char *vpn_type)
{
if (g_strcmp0 (vpn_type, "openvpn") == 0) return "remote";
if (g_strcmp0 (vpn_type, "vpnc") == 0) return "IPSec gateway";
if (g_strcmp0 (vpn_type, "pptp") == 0) return "gateway";
if (g_strcmp0 (vpn_type, "openconnect") == 0) return "gateway";
if (g_strcmp0 (vpn_type, "openswan") == 0) return "right";
cli: extract username and gateway for some more VPN types
2014-05-26 14:51:24 +02:00
if (g_strcmp0 (vpn_type, "libreswan") == 0) return "right";
if (g_strcmp0 (vpn_type, "ssh") == 0) return "remote";
if (g_strcmp0 (vpn_type, "l2tp") == 0) return "gateway";
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
return "";
}
static const gchar *
find_vpn_username_key (const char *vpn_type)
{
if (g_strcmp0 (vpn_type, "openvpn") == 0) return "username";
if (g_strcmp0 (vpn_type, "vpnc") == 0) return "Xauth username";
if (g_strcmp0 (vpn_type, "pptp") == 0) return "user";
if (g_strcmp0 (vpn_type, "openconnect") == 0) return "username";
if (g_strcmp0 (vpn_type, "openswan") == 0) return "leftxauthusername";
cli: extract username and gateway for some more VPN types
2014-05-26 14:51:24 +02:00
if (g_strcmp0 (vpn_type, "libreswan") == 0) return "leftxauthusername";
if (g_strcmp0 (vpn_type, "l2tp") == 0) return "user";
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
return "";
}
enum VpnDataItem {
VPN_DATA_ITEM_GATEWAY,
VPN_DATA_ITEM_USERNAME
};
static const gchar *
get_vpn_data_item (NMConnection *connection, enum VpnDataItem vpn_data_item)
{
const char *key;
char *type = get_vpn_connection_type (connection);
switch (vpn_data_item) {
case VPN_DATA_ITEM_GATEWAY:
key = find_vpn_gateway_key (type);
break;
case VPN_DATA_ITEM_USERNAME:
key = find_vpn_username_key (type);
break;
default:
key = "";
break;
}
g_free (type);
return nm_setting_vpn_get_data_item (nm_connection_get_setting_vpn (connection), key);
}
/* FIXME end */
static gboolean
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
nmc_active_connection_details (NMActiveConnection *acon, NmCli *nmc)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
GError *error = NULL;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
GArray *print_groups;
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
GPtrArray *group_fields = NULL;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
int i;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
const char *fields_str = NULL;
const NMMetaAbstractInfo *const*tmpl;
NmcOutputField *arr;
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
size_t tmpl_len;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
const char *base_hdr = _("Activate connection details");
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
gboolean was_output = FALSE;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
if (!nmc->required_fields || strcasecmp (nmc->required_fields, "common") == 0) {
} else if (!nmc->required_fields || strcasecmp (nmc->required_fields, "all") == 0) {
} else
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
fields_str = nmc->required_fields;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
print_groups = parse_output_fields (fields_str, (const NMMetaAbstractInfo *const*) nmc_fields_con_active_details_groups, TRUE, &group_fields, &error);
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
if (error) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
g_string_printf (nmc->return_text, _("Error: 'connection show': %s"), error->message);
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
g_error_free (error);
cli: add NMC_RESULT_ERROR_USER_INPUT code for bad invocation of nmcli
2010-03-24 13:42:47 +01:00
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
return FALSE;
}
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
g_assert (print_groups);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
/* Main header */
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
{
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
gs_free char *header_name = NULL;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
gs_free NmcOutputField *row = NULL;
gs_unref_array GArray *out_indices = NULL;
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
header_name = construct_header_name (base_hdr, nm_active_connection_get_uuid (acon));
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
out_indices = parse_output_fields (NULL,
(const NMMetaAbstractInfo *const*) nmc_fields_con_active_details_groups,
FALSE, NULL, NULL);
row = g_new0 (NmcOutputField, G_N_ELEMENTS (nmc_fields_con_active_details_groups) + 1);
for (i = 0; nmc_fields_con_active_details_groups[i]; i++)
row[i].info = (const NMMetaAbstractInfo *) nmc_fields_con_active_details_groups[i];
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
cli: merge NmcPrintFields into NmcOutputData and pass it directly to print_required_fields()
2017-03-31 12:55:43 +02:00
print_required_fields (&nmc->nmc_config, NMC_OF_FLAG_MAIN_HEADER_ONLY,
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
out_indices, header_name,
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
0, row);
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
}
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
/* Loop through the groups and print them. */
for (i = 0; i < print_groups->len; i++) {
int group_idx = g_array_index (print_groups, int, i);
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
char *group_fld = (char *) g_ptr_array_index (group_fields, i);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output != NMC_PRINT_TERSE && !nmc->nmc_config.multiline_output && was_output)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("\n"); /* Empty line */
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
was_output = FALSE;
/* GENERAL */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
if (strcasecmp (nmc_fields_con_active_details_groups[group_idx]->name, nmc_fields_con_active_details_groups[0]->name) == 0) {
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
NMC_OUTPUT_DATA_DEFINE_SCOPED (out);
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
/* Add field names */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
tmpl = (const NMMetaAbstractInfo *const*) nmc_fields_con_active_details_general;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
tmpl_len = sizeof (nmc_fields_con_active_details_general);
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
out_indices = parse_output_fields (group_fld,
tmpl, FALSE, NULL, NULL);
arr = nmc_dup_fields_array (tmpl, NMC_OF_FLAG_FIELD_NAMES);
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
g_ptr_array_add (out.output_data, arr);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
/* Fill in values */
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
fill_output_active_connection (acon, out.output_data, TRUE, NMC_OF_FLAG_SECTION_PREFIX);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
print_data_prepare_width (out.output_data);
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
print_data (&nmc->nmc_config, out_indices, NULL, 0, &out);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
was_output = TRUE;
}
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
/* IP4 */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
if (strcasecmp (nmc_fields_con_active_details_groups[group_idx]->name, nmc_fields_con_active_details_groups[1]->name) == 0) {
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
gboolean b1 = FALSE;
libnm: create NMIPConfig as parent of NMIP4Config and NMIP6Config Create NMIPConfig as the parent of NMIP4Config and NMIP6Config, and remove the two subclasses from the public API; while it's convenient to still have both internally, they are now identical to the outside world.
2014-10-22 13:48:18 -04:00
NMIPConfig *cfg4 = nm_active_connection_get_ip4_config (acon);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
b1 = print_ip4_config (cfg4, &nmc->nmc_config, "IP4", group_fld);
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
was_output = was_output || b1;
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
}
/* DHCP4 */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
if (strcasecmp (nmc_fields_con_active_details_groups[group_idx]->name, nmc_fields_con_active_details_groups[2]->name) == 0) {
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
gboolean b1 = FALSE;
libnm: create NMDhcpConfig as parent of NMDhcp4Config and NMDhcp6Config As with NMIP4Config and NMIP6Config, merge the two DHCP config classes into one in the public API.
2014-10-22 13:48:18 -04:00
NMDhcpConfig *dhcp4 = nm_active_connection_get_dhcp4_config (acon);
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
b1 = print_dhcp4_config (dhcp4, &nmc->nmc_config, "DHCP4", group_fld);
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
was_output = was_output || b1;
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
}
/* IP6 */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
if (strcasecmp (nmc_fields_con_active_details_groups[group_idx]->name, nmc_fields_con_active_details_groups[3]->name) == 0) {
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
gboolean b1 = FALSE;
libnm: create NMIPConfig as parent of NMIP4Config and NMIP6Config Create NMIPConfig as the parent of NMIP4Config and NMIP6Config, and remove the two subclasses from the public API; while it's convenient to still have both internally, they are now identical to the outside world.
2014-10-22 13:48:18 -04:00
NMIPConfig *cfg6 = nm_active_connection_get_ip6_config (acon);
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
b1 = print_ip6_config (cfg6, &nmc->nmc_config, "IP6", group_fld);
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
was_output = was_output || b1;
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
}
/* DHCP6 */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
if (strcasecmp (nmc_fields_con_active_details_groups[group_idx]->name, nmc_fields_con_active_details_groups[4]->name) == 0) {
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
gboolean b1 = FALSE;
libnm: create NMDhcpConfig as parent of NMDhcp4Config and NMDhcp6Config As with NMIP4Config and NMIP6Config, merge the two DHCP config classes into one in the public API.
2014-10-22 13:48:18 -04:00
NMDhcpConfig *dhcp6 = nm_active_connection_get_dhcp6_config (acon);
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
b1 = print_dhcp6_config (dhcp6, &nmc->nmc_config, "DHCP6", group_fld);
cli: get IP/DHCP config directly from the NMActiveConnection (rh #1036132) Now that NMActiveConnection has IP and DHCP config info, use that directly in "nmcli con show active" rather than getting it from the connection's device; this way, we get the right info for VPN connections as well.
2013-12-18 11:49:39 -05:00
was_output = was_output || b1;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
}
/* VPN */
if (NM_IS_VPN_CONNECTION (acon) &&
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
strcasecmp (nmc_fields_con_active_details_groups[group_idx]->name, nmc_fields_con_active_details_groups[5]->name) == 0) {
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
NMConnection *con;
NMSettingConnection *s_con;
libnm, core, cli, tui: fix the capitalization of various types GLib/Gtk have mostly settled on the convention that two-letter acronyms in type names remain all-caps (eg, "IO"), but longer acronyms become initial-caps-only (eg, "Tcp"). NM was inconsistent, with most long acronyms using initial caps only (Adsl, Cdma, Dcb, Gsm, Olpc, Vlan), but others using all caps (DHCP, PPP, PPPOE, VPN). Fix libnm and src/ to use initial-caps only for all three-or-more-letter-long acronyms (and update nmcli and nmtui for the libnm changes).
2014-06-26 13:44:36 -04:00
NMSettingVpn *s_vpn;
NMVpnConnectionState vpn_state;
cli: Only escape VPN banner if it's present If the connection is in ACTIVATING state, the banner is still NULL.
2014-11-09 11:36:52 +01:00
char *type_str, *banner_str = NULL, *vpn_state_str;
const char *banner;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
const char *username = NULL;
char **vpn_data_array = NULL;
guint32 items_num;
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
NMC_OUTPUT_DATA_DEFINE_SCOPED (out);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
con = NM_CONNECTION (nm_active_connection_get_connection (acon));
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
s_con = nm_connection_get_setting_connection (con);
g_assert (s_con != NULL);
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
tmpl = (const NMMetaAbstractInfo *const*) nmc_fields_con_active_details_vpn;
out_indices = parse_output_fields (group_fld,
tmpl, FALSE, NULL, NULL);
arr = nmc_dup_fields_array (tmpl, NMC_OF_FLAG_FIELD_NAMES);
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
g_ptr_array_add (out.output_data, arr);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
s_vpn = nm_connection_get_setting_vpn (con);
if (s_vpn) {
items_num = nm_setting_vpn_get_num_data_items (s_vpn);
if (items_num > 0) {
FillVPNDataInfo info;
vpn_data_array = g_new (char *, items_num + 1);
info.array = vpn_data_array;
info.idx = 0;
nm_setting_vpn_foreach_data_item (s_vpn, &fill_vpn_data_item, &info);
vpn_data_array[items_num] = NULL;
}
username = nm_setting_vpn_get_user_name (s_vpn);
}
type_str = get_vpn_connection_type (con);
cli: Only escape VPN banner if it's present If the connection is in ACTIVATING state, the banner is still NULL.
2014-11-09 11:36:52 +01:00
banner = nm_vpn_connection_get_banner (NM_VPN_CONNECTION (acon));
if (banner)
banner_str = g_strescape (banner, "");
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
vpn_state = nm_vpn_connection_get_vpn_state (NM_VPN_CONNECTION (acon));
vpn_state_str = g_strdup_printf ("%d - %s", vpn_state, vpn_connection_state_to_string (vpn_state));
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
/* Add values */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
arr = nmc_dup_fields_array (tmpl, NMC_OF_FLAG_SECTION_PREFIX);
set_val_strc (arr, 0, nmc_fields_con_active_details_groups[5]->name);
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
set_val_str (arr, 1, type_str);
set_val_strc (arr, 2, username ? username : get_vpn_data_item (con, VPN_DATA_ITEM_USERNAME));
set_val_strc (arr, 3, get_vpn_data_item (con, VPN_DATA_ITEM_GATEWAY));
set_val_str (arr, 4, banner_str);
set_val_str (arr, 5, vpn_state_str);
set_val_arr (arr, 6, vpn_data_array);
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
g_ptr_array_add (out.output_data, arr);
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
print_data_prepare_width (out.output_data);
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
print_data (&nmc->nmc_config, out_indices, NULL, 0, &out);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
was_output = TRUE;
}
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
}
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
g_array_free (print_groups, TRUE);
cli: allow '--field group.field' for nmcli con show active <bla> Example: $ nmcli -f general.name,ip4.address c s a myeth GENERAL.NAME: myeth IP4.ADDRESS[1]: ip = 10.34.25.228/23, gw = 10.34.25.254
2013-12-10 15:33:53 +01:00
if (group_fields)
g_ptr_array_free (group_fields, TRUE);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
return TRUE;
}
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
static gboolean
split_required_fields_for_con_show (const char *input,
char **profile_flds,
char **active_flds,
GError **error)
{
char **fields, **iter;
char *dot;
GString *str1, *str2;
gboolean found;
gboolean group_profile = FALSE;
gboolean group_active = FALSE;
gboolean success = TRUE;
gboolean is_all, is_common;
int i;
if (!input) {
*profile_flds = NULL;
*active_flds = NULL;
return TRUE;
}
str1 = g_string_new (NULL);
str2 = g_string_new (NULL);
/* Split supplied fields string */
fields = g_strsplit_set (input, ",", -1);
for (iter = fields; iter && *iter; iter++) {
g_strstrip (*iter);
dot = strchr (*iter, '.');
if (dot)
*dot = '\0';
is_all = !dot && strcasecmp (*iter, "all") == 0;
is_common = !dot && strcasecmp (*iter, "common") == 0;
found = FALSE;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
for (i = 0; i < _NM_META_SETTING_TYPE_NUM; i++) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
if ( is_all || is_common
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
|| !strcasecmp (*iter, nm_meta_setting_infos[i].setting_name)) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
if (dot)
*dot = '.';
g_string_append (str1, *iter);
g_string_append_c (str1, ',');
found = TRUE;
break;
}
}
if (found)
continue;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
for (i = 0; nmc_fields_con_active_details_groups[i]; i++) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
if ( is_all || is_common
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
|| !strcasecmp (*iter, nmc_fields_con_active_details_groups[i]->name)) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
if (dot)
*dot = '.';
g_string_append (str2, *iter);
g_string_append_c (str2, ',');
found = TRUE;
break;
}
}
if (!found) {
if (dot)
*dot = '.';
if (!strcasecmp (*iter, CON_SHOW_DETAIL_GROUP_PROFILE))
group_profile = TRUE;
else if (!strcasecmp (*iter, CON_SHOW_DETAIL_GROUP_ACTIVE))
group_active = TRUE;
else {
cli: implement new nmc_print() command to generically output cli data We already have - data sources (nm_cli, connections or settings) - meta data information how to access the data sources (NMMetaAbstractInfo, NmcMetaGenericInfo, NMMetaPropertyInfo) Add now a generic way to output cli data using nmc_print(). It gets a list of data-sources (@targets) and a list of available fields (meta data). It also gets cli configuration (NmcConfig) and field selector strings (@field_str). Based on that, it should output the desired data. This is intended to replaces the previous approach, where functions like show_nm_status() have full knowledge about how to access the data and create an intermediate output format (NmcOutputData, NmcOutputField) that was printed via print_data(). show_nm_status() contained both knowledge about the data itself (how to print a value) and intimate knoweledge about the output intermediate format. Also, the intermediate format is hard to understand. For example, sometimes we put the field prefix in NmcOutputField at index 0 and via the NmcOfFlags we control how to output the data. Clearly separate the responsibilities. - The meta data (NmcMetaGenericInfo) is only concerned with converting a data source to a string (or a color format). - the field selection (@field_str) only cares about parsing the list of NMMetaAbstractInfo. - _print_fill() populates a table with output values and header entries. - _print_do() prints the previously prepared table. The advantage is that if you want to change anything, you only need to touch a particular part. This is only a show-case for `nmcli general status`. Parts are still un-implemented and will follow. This changes behavior for --terse mode: the values are now no longer translated: $ LANG=de_DE.utf8 nmcli -t --mode multiline general
2017-04-04 15:23:39 +02:00
char *allowed1 = nmc_get_allowed_fields ((const NMMetaAbstractInfo *const*) nm_meta_setting_infos_editor_p (), NULL);
char *allowed2 = nmc_get_allowed_fields ((const NMMetaAbstractInfo *const*) nmc_fields_con_active_details_groups, NULL);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
g_set_error (error, NMCLI_ERROR, 0, _("invalid field '%s'; allowed fields: %s and %s, or %s,%s"),
*iter, allowed1, allowed2, CON_SHOW_DETAIL_GROUP_PROFILE, CON_SHOW_DETAIL_GROUP_ACTIVE);
g_free (allowed1);
g_free (allowed2);
success = FALSE;
break;
}
}
}
if (fields)
g_strfreev (fields);
/* Handle pseudo groups: profile, active */
if (success && group_profile) {
if (str1->len > 0) {
fix typos in documentation and messages https://bugzilla.gnome.org/show_bug.cgi?id=727031 Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-03-25 18:40:47 +02:00
g_set_error (error, NMCLI_ERROR, 0, _("'%s' has to be alone"),
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
CON_SHOW_DETAIL_GROUP_PROFILE);
success = FALSE;
} else
g_string_assign (str1, "all,");
}
if (success && group_active) {
if (str2->len > 0) {
fix typos in documentation and messages https://bugzilla.gnome.org/show_bug.cgi?id=727031 Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-03-25 18:40:47 +02:00
g_set_error (error, NMCLI_ERROR, 0, _("'%s' has to be alone"),
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
CON_SHOW_DETAIL_GROUP_ACTIVE);
success = FALSE;
} else
g_string_assign (str2, "all,");
}
if (success) {
if (str1->len > 0)
g_string_truncate (str1, str1->len - 1);
if (str2->len > 0)
g_string_truncate (str2, str2->len - 1);
*profile_flds = g_string_free (str1, str1->len == 0);
*active_flds = g_string_free (str2, str2->len == 0);
} else {
g_string_free (str1, TRUE);
g_string_free (str2, TRUE);
}
return success;
}
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
typedef enum {
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
NMC_SORT_ACTIVE = 1,
NMC_SORT_ACTIVE_INV = -1,
NMC_SORT_NAME = 2,
NMC_SORT_NAME_INV = -2,
NMC_SORT_TYPE = 3,
NMC_SORT_TYPE_INV = -3,
NMC_SORT_PATH = 4,
NMC_SORT_PATH_INV = -4,
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
} NmcSortOrder;
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
typedef struct {
NmCli *nmc;
const GArray *order;
} NmcSortInfo;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
static int
compare_connections (gconstpointer a, gconstpointer b, gpointer user_data)
{
NMConnection *ca = *(NMConnection **)a;
NMConnection *cb = *(NMConnection **)b;
NMActiveConnection *aca, *acb;
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
NmcSortInfo *info = (NmcSortInfo *) user_data;
GArray *default_order = NULL;
const GArray *order;
NmcSortOrder item;
int cmp = 0, i;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
const char *tmp1, *tmp2;
unsigned long tmp1_int, tmp2_int;
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
if (info->order )
order = info->order;
else {
NmcSortOrder def[] = { NMC_SORT_ACTIVE, NMC_SORT_NAME, NMC_SORT_PATH };
int num = G_N_ELEMENTS (def);
default_order = g_array_sized_new (FALSE, FALSE, sizeof (NmcSortOrder), num);
g_array_append_vals (default_order, def, num);
order = default_order;
}
for (i = 0; i < order->len; i++) {
cli/connections: split out ac_state_to_color() Going to be useful outside nmcli c.
2016-07-01 21:44:47 +02:00
item = g_array_index (order, NmcSortOrder, i);
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
switch (item) {
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
case NMC_SORT_ACTIVE:
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
case NMC_SORT_ACTIVE_INV:
aca = get_ac_for_connection (nm_client_get_active_connections (info->nmc->client), ca);
acb = get_ac_for_connection (nm_client_get_active_connections (info->nmc->client), cb);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
cmp = (aca && !acb) ? -1 : (!aca && acb) ? 1 : 0;
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
if (item == NMC_SORT_ACTIVE_INV)
cmp = -(cmp);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
break;
case NMC_SORT_TYPE:
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
case NMC_SORT_TYPE_INV:
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
cmp = g_strcmp0 (nm_connection_get_connection_type (ca),
nm_connection_get_connection_type (cb));
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
if (item == NMC_SORT_TYPE_INV)
cmp = -(cmp);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
break;
case NMC_SORT_NAME:
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
case NMC_SORT_NAME_INV:
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
cmp = g_strcmp0 (nm_connection_get_id (ca),
nm_connection_get_id (cb));
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
if (item == NMC_SORT_NAME_INV)
cmp = -(cmp);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
break;
case NMC_SORT_PATH:
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
case NMC_SORT_PATH_INV:
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
tmp1 = nm_connection_get_path (ca);
tmp2 = nm_connection_get_path (cb);
tmp1 = tmp1 ? strrchr (tmp1, '/') : "0";
tmp2 = tmp2 ? strrchr (tmp2, '/') : "0";
nmc_string_to_uint (tmp1 ? tmp1+1 : "0", FALSE, 0, 0, &tmp1_int);
nmc_string_to_uint (tmp2 ? tmp2+1 : "0", FALSE, 0, 0, &tmp2_int);
cmp = (int) tmp1_int - tmp2_int;
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
if (item == NMC_SORT_PATH_INV)
cmp = -(cmp);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
break;
default:
cmp = 0;
break;
}
if (cmp != 0)
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
goto end;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
}
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
end:
if (default_order)
g_array_unref (default_order);
return cmp;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
}
static GPtrArray *
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
sort_connections (const GPtrArray *cons, NmCli *nmc, const GArray *order)
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
{
GPtrArray *sorted;
int i;
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
NmcSortInfo compare_info;
client: fix NULL dereference with no connections Coverity complains: CID 59386 (#1 of 1): Dereference before null check (REVERSE_INULL) check_after_deref: Null-checking cons suggests that it may be null, but it has already been dereferenced on all paths leading to the check.
2016-02-04 16:47:00 +01:00
if (!cons)
cli: fix compile error connections.c: In function ‘sort_connections’: connections.c:1536:3: error: ‘return’ with no value, in function returning non-void [-Werror=return-type] return; ^ Fixes: e8927c14ec3e830099e1525fa8fee4c9a7fc5032
2016-02-11 17:14:35 +01:00
return NULL;
client: fix NULL dereference with no connections Coverity complains: CID 59386 (#1 of 1): Dereference before null check (REVERSE_INULL) check_after_deref: Null-checking cons suggests that it may be null, but it has already been dereferenced on all paths leading to the check.
2016-02-04 16:47:00 +01:00
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
compare_info.nmc = nmc;
compare_info.order = order;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
sorted = g_ptr_array_sized_new (cons->len);
client: fix NULL dereference with no connections Coverity complains: CID 59386 (#1 of 1): Dereference before null check (REVERSE_INULL) check_after_deref: Null-checking cons suggests that it may be null, but it has already been dereferenced on all paths leading to the check.
2016-02-04 16:47:00 +01:00
for (i = 0; i < cons->len; i++)
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
g_ptr_array_add (sorted, cons->pdata[i]);
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
g_ptr_array_sort_with_data (sorted, compare_connections, &compare_info);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
return sorted;
}
static int
compare_ac_connections (gconstpointer a, gconstpointer b, gpointer user_data)
{
NMActiveConnection *ca = *(NMActiveConnection **)a;
NMActiveConnection *cb = *(NMActiveConnection **)b;
int cmp;
/* Sort states first */
cmp = nm_active_connection_get_state (cb) - nm_active_connection_get_state (ca);
if (cmp != 0)
return cmp;
cmp = g_strcmp0 (nm_active_connection_get_id (ca),
nm_active_connection_get_id (cb));
if (cmp != 0)
return cmp;
return g_strcmp0 (nm_active_connection_get_connection_type (ca),
nm_active_connection_get_connection_type (cb));
}
static GPtrArray *
get_invisible_active_connections (NmCli *nmc)
{
const GPtrArray *acons;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
GPtrArray *invisibles;
int a, c;
g_return_val_if_fail (nmc != NULL, NULL);
invisibles = g_ptr_array_new ();
acons = nm_client_get_active_connections (nmc->client);
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
for (a = 0; a < acons->len; a++) {
gboolean found = FALSE;
NMActiveConnection *acon = g_ptr_array_index (acons, a);
const char *a_uuid = nm_active_connection_get_uuid (acon);
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
for (c = 0; c < connections->len; c++) {
NMConnection *con = g_ptr_array_index (connections, c);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
const char *c_uuid = nm_connection_get_uuid (con);
if (strcmp (a_uuid, c_uuid) == 0) {
found = TRUE;
break;
}
}
/* Active connection is not in connections array, add it to */
if (!found)
g_ptr_array_add (invisibles, acon);
}
g_ptr_array_sort_with_data (invisibles, compare_ac_connections, NULL);
return invisibles;
}
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
static GArray *
parse_preferred_connection_order (const char *order, GError **error)
{
char **strv, **iter;
const char *str;
GArray *order_arr;
NmcSortOrder val;
gboolean inverse, unique;
int i;
strv = nmc_strsplit_set (order, ":", -1);
if (!strv || !*strv) {
g_set_error (error, NMCLI_ERROR, 0,
_("incorrect string '%s' of '--order' option"), order);
g_strfreev (strv);
return NULL;
}
order_arr = g_array_sized_new (FALSE, FALSE, sizeof (NmcSortOrder), 4);
for (iter = strv; iter && *iter; iter++) {
str = *iter;
inverse = FALSE;
if (str[0] == '-')
inverse = TRUE;
if (str[0] == '+' || str[0] == '-')
str++;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (str, "active"))
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
val = inverse ? NMC_SORT_ACTIVE_INV : NMC_SORT_ACTIVE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (str, "name"))
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
val = inverse ? NMC_SORT_NAME_INV : NMC_SORT_NAME;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (str, "type"))
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
val = inverse ? NMC_SORT_TYPE_INV : NMC_SORT_TYPE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (str, "path"))
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
val = inverse ? NMC_SORT_PATH_INV : NMC_SORT_PATH;
else {
g_array_unref (order_arr);
order_arr = NULL;
g_set_error (error, NMCLI_ERROR, 0,
_("incorrect item '%s' in '--order' option"), *iter);
break;
}
/* Check for duplicates and ignore them. */
unique = TRUE;
for (i = 0; i < order_arr->len; i++) {
if (abs (g_array_index (order_arr, NmcSortOrder, i)) - abs (val) == 0) {
unique = FALSE;
break;
}
}
/* Value is ok and unique, add it to the array */
if (unique)
g_array_append_val (order_arr, val);
}
g_strfreev (strv);
return order_arr;
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
static NMConnection *
get_connection (NmCli *nmc, int *argc, char ***argv, int *pos, GError **error)
{
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
NMConnection *connection = NULL;
const char *selector = NULL;
if (*argc == 0) {
g_set_error_literal (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("No connection specified"));
return NULL;
}
cli/connections: do connection completion in get_connection() Start completing by the id if the filter type is not specified
2016-07-07 14:53:34 +02:00
if (*argc == 1 && nmc->complete)
nmc_complete_strings (**argv, "id", "uuid", "path", NULL);
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
if ( strcmp (**argv, "id") == 0
|| strcmp (**argv, "uuid") == 0
|| strcmp (**argv, "path") == 0) {
selector = **argv;
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
(*argc)--;
(*argv)++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!*argc) {
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("%s argument is missing"), selector);
return NULL;
}
}
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
connection = nmc_find_connection (connections, selector, **argv, pos,
cli/connections: do connection completion in get_connection() Start completing by the id if the filter type is not specified
2016-07-07 14:53:34 +02:00
*argc == 1 && nmc->complete);
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
if (!connection) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_NOT_FOUND,
_("unknown connection '%s'"), **argv);
}
cli: fix connection delete/monitor The two commands should apply to all available connections matching the given name. Fixes: 717db4fe014ecc80eb8bb86757f724448116a662
2016-08-02 16:49:37 +02:00
/* If the caller wants multiple results (pos is set) and there are any,
* don't switch to next argument.
*/
if (!pos || !*pos)
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, argc, argv, NULL);
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
return connection;
}
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
static NMCResultCode
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
do_connections_show (NmCli *nmc, int argc, char **argv)
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
{
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
GError *err = NULL;
char *profile_flds = NULL, *active_flds = NULL;
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
GPtrArray *invisibles, *sorted_cons;
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
gboolean active_only = FALSE;
GArray *order = NULL;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
int i, option;
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
/* check connection show options [--active] [--order <order spec>] */
while ((option = next_arg (nmc, &argc, &argv, "--active", "--order", NULL)) > 0) {
switch (option) {
case 1: /* --active */
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
active_only = TRUE;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
break;
case 2: /* --order */
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
g_set_error_literal (&err, NMCLI_ERROR, 0,
_("'--order' argument is missing"));
goto finish;
}
order = parse_preferred_connection_order (*argv, &err);
if (err)
goto finish;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
break;
default:
g_assert_not_reached();
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
break;
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
}
}
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
if (argc == 0) {
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
const char *fields_str = NULL;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
char *fields_common = NMC_FIELDS_CON_SHOW_COMMON;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
const NMMetaAbstractInfo *const*tmpl;
NmcOutputField *arr;
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
NMC_OUTPUT_DATA_DEFINE_SCOPED (out);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
if (nmc->complete)
goto finish;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
if (!nmc->required_fields || strcasecmp (nmc->required_fields, "common") == 0)
fields_str = fields_common;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
else if (!nmc->required_fields || strcasecmp (nmc->required_fields, "all") == 0) {
} else
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
fields_str = nmc->required_fields;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
tmpl = (const NMMetaAbstractInfo *const*) nmc_fields_con_show;
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
out_indices = parse_output_fields (fields_str, tmpl, FALSE, NULL, &err);
nmcli: don't enforce anymore the -f(ields) option in -t(erse) mode
2017-01-16 13:54:33 +01:00
if (err)
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
goto finish;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
cli: adapt the code for the new flags and printing facilities The data are added to output_data at first, and then they are printed all at once using print_data(), that takes care of proper alignment and display. The static 'width' values defined in NmcOutputField columns are not used now, but dynamically computed maximal widths override them.
2013-05-22 08:37:50 +02:00
/* Add headers */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
arr = nmc_dup_fields_array (tmpl, NMC_OF_FLAG_MAIN_HEADER_ADD | NMC_OF_FLAG_FIELD_NAMES);
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
g_ptr_array_add (out.output_data, arr);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
/* There might be active connections not present in connection list
* (e.g. private connections of a different user). Show them as well. */
invisibles = get_invisible_active_connections (nmc);
for (i = 0; i < invisibles->len; i++)
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
fill_output_connection_for_invisible (invisibles->pdata[i], out.output_data);
cli: fix memory leak in connection listing
2016-03-14 15:05:28 +01:00
g_ptr_array_free (invisibles, TRUE);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
/* Sort the connections and fill the output data */
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
sorted_cons = sort_connections (connections, nmc, order);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
for (i = 0; i < sorted_cons->len; i++)
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
fill_output_connection (sorted_cons->pdata[i], nmc->client, out.output_data, active_only);
cli: fix memory leak in connection listing
2016-03-14 15:05:28 +01:00
g_ptr_array_free (sorted_cons, TRUE);
nmcli: sort connections in 'nmcli connection show' output
2015-02-12 17:31:06 +01:00
cli: don't track output data in global NmCli structure We should not violate the global data to track the output data while it is constructed and printed. Most of the time, we actually clear the output data anyway -- either before constructing it, or after printing it. In some cases we didn't, but I think that is a bug. It's really hard to keep track of this. The output data should belong to a certain scope and get destroyed afterwards. Passing it around is very confusing. Don't do that.
2017-03-30 14:56:19 +02:00
print_data_prepare_width (out.output_data);
cli: pass arguments for print_data separately of NmcOutputData Don't pass on large structs of input arguments. It only convolutes which arguments are passed, and where they come from.
2017-03-31 13:21:47 +02:00
print_data (&nmc->nmc_config, out_indices,
active_only ? _("NetworkManager active profiles")
: _("NetworkManager connection profiles"),
0, &out);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
} else {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
gboolean new_line = FALSE;
gboolean without_fields = (nmc->required_fields == NULL);
const GPtrArray *active_cons = nm_client_get_active_connections (nmc->client);
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
int pos = 0;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
/* multiline mode is default for 'connection show <ID>' */
if (!nmc->mode_specified)
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
nmc->nmc_config_mutable.multiline_output = TRUE;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
/* Split required fields into the settings and active ones. */
if (!split_required_fields_for_con_show (nmc->required_fields, &profile_flds, &active_flds, &err))
goto finish;
g_free (nmc->required_fields);
nmc->required_fields = NULL;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
while (argc > 0) {
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
gboolean res;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
NMConnection *con;
NMActiveConnection *acon = NULL;
cli: make id|uuid specifiers optional for 'connection list' and 'connection status' And also allow: - identifying connetions with 'path'/'apath' accepting either the whole D-Dus path or just an index - multiple connections on cmdline at once nmcli connection list [[id|uuid|path] <ID>] nmcli connection status [[id|uuid|path|apath] <ID>] Examples: nmcli c l nmcli c l "My Wi-Fi 1" airpoirt1 cpath 23 id "My Wi-Fi 2" nmcli c s nmcli c s apath 1 ee19b893-e2a7-42cf-ac82-34506f07a05f
2012-11-16 12:43:44 +01:00
const char *selector = NULL;
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
if (argc == 1 && nmc->complete)
nmc_complete_strings (*argv, "id", "uuid", "path", "apath", NULL);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
if ( strcmp (*argv, "id") == 0
|| strcmp (*argv, "uuid") == 0
cli: make id|uuid specifiers optional for 'connection list' and 'connection status' And also allow: - identifying connetions with 'path'/'apath' accepting either the whole D-Dus path or just an index - multiple connections on cmdline at once nmcli connection list [[id|uuid|path] <ID>] nmcli connection status [[id|uuid|path|apath] <ID>] Examples: nmcli c l nmcli c l "My Wi-Fi 1" airpoirt1 cpath 23 id "My Wi-Fi 2" nmcli c s nmcli c s apath 1 ee19b893-e2a7-42cf-ac82-34506f07a05f
2012-11-16 12:43:44 +01:00
|| strcmp (*argv, "path") == 0
|| strcmp (*argv, "apath") == 0) {
selector = *argv;
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
cli: allow next_arg() to move arguments past the last one
2012-11-16 22:43:29 +01:00
g_string_printf (nmc->return_text, _("Error: %s argument is missing."), *(argv-1));
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
goto finish;
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
}
cli: make id|uuid specifiers optional for 'connection list' and 'connection status' And also allow: - identifying connetions with 'path'/'apath' accepting either the whole D-Dus path or just an index - multiple connections on cmdline at once nmcli connection list [[id|uuid|path] <ID>] nmcli connection status [[id|uuid|path|apath] <ID>] Examples: nmcli c l nmcli c l "My Wi-Fi 1" airpoirt1 cpath 23 id "My Wi-Fi 2" nmcli c s nmcli c s apath 1 ee19b893-e2a7-42cf-ac82-34506f07a05f
2012-11-16 12:43:44 +01:00
}
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
/* Try to find connection by id, uuid or path first */
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
con = nmc_find_connection (connections, selector, *argv, &pos,
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
argc == 1 && nmc->complete);
if (!con && (!selector || strcmp (selector, "apath") == 0)) {
/* Try apath too */
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
acon = find_active_connection (active_cons, connections, "apath", *argv, NULL,
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
argc == 1 && nmc->complete);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
if (acon)
libnm: fix NMActiveConnection object-path-valued properties NMActiveConnection:connection was an object path rather than an NMRemoteConnection because in the past the NMObject property system wasn't capable of handling NMRemoteConnection-valued properties. NMActiveConnection:master was an object path rather than an NMDevice entirely by accident. Fix both. NMActiveConnection:specific-object can't currently be converted to an object, because we don't know ahead of time what object type it is, and NMObject can't deal with that. Instead, we rename it to :specific-object-path (and likewise for its get function), both to emphasize that it doesn't behave like other properties, and to leave the old name open for an actual object-valued property later on.
2014-10-09 09:24:05 -04:00
con = NM_CONNECTION (nm_active_connection_get_connection (acon));
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
}
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
if (!con && !acon) {
g_string_printf (nmc->return_text, _("Error: %s - no such connection profile."), *argv);
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
goto finish;
}
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
/* Print connection details:
* Usually we have both static and active connection.
* But when a connection is private to a user, another user
* may see only the active connection.
*/
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
/* Filter only active connections */
if (!acon)
acon = get_ac_for_connection (active_cons, con);
if (active_only && !acon) {
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
continue;
}
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
if (nmc->complete) {
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli/connections: add completion to show subcommand
2016-07-11 22:38:16 +02:00
continue;
}
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
/* Show an empty line between connections */
if (new_line)
g_print ("\n");
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
/* Show profile configuration */
if (without_fields || profile_flds) {
if (con) {
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
nmc->required_fields = profile_flds;
cli/connections: drop the redundant parsing of --show-secrets This is no longer needed now that --show-secrets is a global option.
2017-03-24 14:36:12 +01:00
if (nmc->show_secrets)
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
update_secrets_in_connection (NM_REMOTE_CONNECTION (con), con);
cli/connections: drop the redundant parsing of --show-secrets This is no longer needed now that --show-secrets is a global option.
2017-03-24 14:36:12 +01:00
res = nmc_connection_profile_details (con, nmc, nmc->show_secrets);
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
nmc->required_fields = NULL;
if (!res)
goto finish;
}
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
}
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
/* If the profile is active, print also active details */
if (without_fields || active_flds) {
if (acon) {
nmc->required_fields = active_flds;
res = nmc_active_connection_details (acon, nmc);
nmc->required_fields = NULL;
if (!res)
goto finish;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
}
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
}
cli: show details for active connection even if the profile is not visible $ nmcli con show other_user_con
2014-10-16 17:51:59 +02:00
new_line = TRUE;
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
/* Take next argument.
* But for pos != NULL we have more connections of the same name,
* so process the same argument again.
*/
if (!pos)
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
}
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
finish:
if (err) {
g_string_printf (nmc->return_text, _("Error: %s."), err->message);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
g_error_free (err);
cli: added 'con status id|uuid|path <con>' for listing active connection details 'nmcli con status' still prints brief active connection list. 'nmcli con status id|uuid|path <con>' shows details of the selected active connection.
2012-01-06 15:15:03 +01:00
}
cli: consolidate active and configured connections (rh #997999) Handle connection profiles in a single 'show' command instead of 'show active' and 'show configured'. nmcli con show [--active] [[id|uuid|path|apath] <bla>] nmcli con show : display all connection profiles nmcli con show --active : only display active connection profiles (filters out inactive profiles) nmcli con show myeth : display details of "myeth" profile, and also active connection info (if the profile is active) nmcli -f profile con show myeth : only display "myeth"'s static configuration nmcli -f active con show myeth : only display active details of "myeth" nmcli -f connection.id,ipv4,general con show myeth : display "connection.id"a property "ipv4" setting and "GENERAL" group of active data https://bugzilla.redhat.com/show_bug.cgi?id=997999
2013-11-07 14:47:12 +01:00
g_free (profile_flds);
g_free (active_flds);
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
if (order)
g_array_unref (order);
cli: changing nmcli output to better fit both computer and human needs The output is basically tabular with fields (columns) presenting specific pieces of info. Each line represents a single object. It's possible to switch to multiline output using '--multiline' option. In that mode single object is presented on more lines - each field on its line. Terse mode now uses ':' as field separator. It also escapes all occurences of ':' and '\' inside field values to ease parsing. The escaping behaviour can be controlled through '--escape' option. By default, escaping is switched on in tabular mode. When using terse mode ('--terse'), '--fields' option is mandatory for specifying required fields. That helps for flexibility and backwards compatibility. Not all output is converted yet.
2010-03-18 15:39:15 +01:00
return nmc->return_value;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
static NMActiveConnection *
get_default_active_connection (NmCli *nmc, NMDevice **device)
{
NMActiveConnection *default_ac = NULL;
NMDevice *non_default_device = NULL;
NMActiveConnection *non_default_ac = NULL;
const GPtrArray *connections;
int i;
g_return_val_if_fail (nmc != NULL, NULL);
g_return_val_if_fail (device != NULL, NULL);
g_return_val_if_fail (*device == NULL, NULL);
connections = nm_client_get_active_connections (nmc->client);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (i = 0; i < connections->len; i++) {
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
NMActiveConnection *candidate = g_ptr_array_index (connections, i);
const GPtrArray *devices;
devices = nm_active_connection_get_devices (candidate);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
if (!devices->len)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
continue;
if (nm_active_connection_get_default (candidate)) {
if (!default_ac) {
*device = g_ptr_array_index (devices, 0);
default_ac = candidate;
}
} else {
if (!non_default_ac) {
non_default_device = g_ptr_array_index (devices, 0);
non_default_ac = candidate;
}
}
}
/* Prefer the default connection if one exists, otherwise return the first
* non-default connection.
*/
if (!default_ac && non_default_ac) {
default_ac = non_default_ac;
*device = non_default_device;
}
return default_ac;
}
/* Find a device to activate the connection on.
* IN: connection: connection to activate
* iface: device interface name to use (optional)
* ap: access point to use (optional; valid just for 802-11-wireless)
cli: add support for WiMAX devices and connections
2011-01-06 17:01:55 -06:00
* nsp: Network Service Provider to use (option; valid only for wimax)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
* OUT: device: found device
* spec_object: specific_object path of NMAccessPoint
* RETURNS: TRUE when a device is found, FALSE otherwise.
*/
static gboolean
cli: add support for WiMAX devices and connections
2011-01-06 17:01:55 -06:00
find_device_for_connection (NmCli *nmc,
NMConnection *connection,
const char *iface,
const char *ap,
const char *nsp,
NMDevice **device,
const char **spec_object,
GError **error)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
NMSettingConnection *s_con;
const char *con_type;
int i, j;
g_return_val_if_fail (nmc != NULL, FALSE);
cli: don't look up a device for activation request unless we have to Let the server decide which device to use if the user didn't explicitly specify the interface, wireless access point or a wimax nsp. The server will just reuse the device for an already active connection or potentially do a better guess. https://bugzilla.gnome.org/show_bug.cgi?id=730492
2015-02-01 09:09:37 +01:00
g_return_val_if_fail (iface || ap || nsp, FALSE);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
g_return_val_if_fail (device != NULL && *device == NULL, FALSE);
g_return_val_if_fail (spec_object != NULL && *spec_object == NULL, FALSE);
g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
cli: use setting accessor helpers
2011-03-16 14:32:24 +01:00
s_con = nm_connection_get_setting_connection (connection);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
g_assert (s_con);
con_type = nm_setting_connection_get_connection_type (s_con);
cli: use libnm-util's constants for settings types
2011-09-06 16:04:12 +02:00
if (strcmp (con_type, NM_SETTING_VPN_SETTING_NAME) == 0) {
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
/* VPN connections */
NMActiveConnection *active = NULL;
if (iface) {
cli: simplify find_device_for_connection()
2011-06-15 15:00:55 +02:00
*device = nm_client_get_device_by_iface (nmc->client, iface);
if (*device)
active = nm_device_get_active_connection (*device);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
if (!active) {
cli: use an error quark for domain in g_set_error() instead of 0 (rh #842975) glib 2.32 makes a runtime check that domain is not NULL: GLib-WARNING **: (gerror.c:390):g_error_new_valist: runtime check failed: (domain != 0)
2012-07-25 13:57:45 +02:00
g_set_error (error, NMCLI_ERROR, 0, _("no active connection on device '%s'"), iface);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
return FALSE;
}
*spec_object = nm_object_get_path (NM_OBJECT (active));
return TRUE;
} else {
active = get_default_active_connection (nmc, device);
if (!active) {
cli: use an error quark for domain in g_set_error() instead of 0 (rh #842975) glib 2.32 makes a runtime check that domain is not NULL: GLib-WARNING **: (gerror.c:390):g_error_new_valist: runtime check failed: (domain != 0)
2012-07-25 13:57:45 +02:00
g_set_error_literal (error, NMCLI_ERROR, 0, _("no active connection or device"));
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
return FALSE;
}
*spec_object = nm_object_get_path (NM_OBJECT (active));
return TRUE;
}
} else {
/* Other connections */
NMDevice *found_device = NULL;
const GPtrArray *devices = nm_client_get_devices (nmc->client);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (i = 0; i < devices->len && !found_device; i++) {
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
NMDevice *dev = g_ptr_array_index (devices, i);
if (iface) {
const char *dev_iface = nm_device_get_iface (dev);
cli: improve devices vs connection compatibility check report error message on device compatibility failures
2016-07-15 14:38:00 +02:00
if (!nm_streq0 (dev_iface, iface))
continue;
if (!nm_device_connection_compatible (dev, connection, error)) {
g_prefix_error (error, _("device '%s' not compatible with connection '%s':"),
iface, nm_setting_connection_get_id (s_con));
return FALSE;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: improve devices vs connection compatibility check report error message on device compatibility failures
2016-07-15 14:38:00 +02:00
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
} else {
cli: improve devices vs connection compatibility check report error message on device compatibility failures
2016-07-15 14:38:00 +02:00
if (!nm_device_connection_compatible (dev, connection, NULL))
continue;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: improve devices vs connection compatibility check report error message on device compatibility failures
2016-07-15 14:38:00 +02:00
found_device = dev;
if (ap && !strcmp (con_type, NM_SETTING_WIRELESS_SETTING_NAME) && NM_IS_DEVICE_WIFI (dev)) {
cli: use nm_access_point_get_bssid() instead of nm_access_point_get_hw_address() nm_access_point_get_hw_address() is deprecated in 0.9. Also change 'hwaddr' parameter for 'nmcli dev wifi list' to 'bssid'. 'hwaddr' still works but is deprecated and not documented any more.
2011-09-06 13:37:23 +02:00
char *bssid_up = g_ascii_strup (ap, -1);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
const GPtrArray *aps = nm_device_wifi_get_access_points (NM_DEVICE_WIFI (dev));
found_device = NULL; /* Mark as not found; set to the device again later, only if AP matches */
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (j = 0; j < aps->len; j++) {
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
NMAccessPoint *candidate_ap = g_ptr_array_index (aps, j);
cli: use nm_access_point_get_bssid() instead of nm_access_point_get_hw_address() nm_access_point_get_hw_address() is deprecated in 0.9. Also change 'hwaddr' parameter for 'nmcli dev wifi list' to 'bssid'. 'hwaddr' still works but is deprecated and not documented any more.
2011-09-06 13:37:23 +02:00
const char *candidate_bssid = nm_access_point_get_bssid (candidate_ap);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: use nm_access_point_get_bssid() instead of nm_access_point_get_hw_address() nm_access_point_get_hw_address() is deprecated in 0.9. Also change 'hwaddr' parameter for 'nmcli dev wifi list' to 'bssid'. 'hwaddr' still works but is deprecated and not documented any more.
2011-09-06 13:37:23 +02:00
if (!strcmp (bssid_up, candidate_bssid)) {
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
found_device = dev;
*spec_object = nm_object_get_path (NM_OBJECT (candidate_ap));
break;
}
}
cli: use nm_access_point_get_bssid() instead of nm_access_point_get_hw_address() nm_access_point_get_hw_address() is deprecated in 0.9. Also change 'hwaddr' parameter for 'nmcli dev wifi list' to 'bssid'. 'hwaddr' still works but is deprecated and not documented any more.
2011-09-06 13:37:23 +02:00
g_free (bssid_up);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: add support for WiMAX devices and connections
2011-01-06 17:01:55 -06:00
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
if (found_device) {
*device = found_device;
return TRUE;
} else {
if (iface)
cli: use an error quark for domain in g_set_error() instead of 0 (rh #842975) glib 2.32 makes a runtime check that domain is not NULL: GLib-WARNING **: (gerror.c:390):g_error_new_valist: runtime check failed: (domain != 0)
2012-07-25 13:57:45 +02:00
g_set_error (error, NMCLI_ERROR, 0, _("device '%s' not compatible with connection '%s'"),
iface, nm_setting_connection_get_id (s_con));
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
else
cli: use an error quark for domain in g_set_error() instead of 0 (rh #842975) glib 2.32 makes a runtime check that domain is not NULL: GLib-WARNING **: (gerror.c:390):g_error_new_valist: runtime check failed: (domain != 0)
2012-07-25 13:57:45 +02:00
g_set_error (error, NMCLI_ERROR, 0, _("no device found for connection '%s'"),
nm_setting_connection_get_id (s_con));
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
return FALSE;
}
}
}
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
typedef struct {
NmCli *nmc;
NMDevice *device;
NMActiveConnection *active;
} ActivateConnectionInfo;
static void activate_connection_info_finish (ActivateConnectionInfo *info);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
static const char *
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
active_connection_state_reason_to_string (NMActiveConnectionStateReason reason)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
switch (reason) {
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
case NM_ACTIVE_CONNECTION_STATE_REASON_UNKNOWN:
return _("Unknown reason");
case NM_ACTIVE_CONNECTION_STATE_REASON_NONE:
return _("The connection was disconnected");
case NM_ACTIVE_CONNECTION_STATE_REASON_USER_DISCONNECTED:
return _("Disconnected by user");
case NM_ACTIVE_CONNECTION_STATE_REASON_DEVICE_DISCONNECTED:
return _("The base network connection was interrupted");
case NM_ACTIVE_CONNECTION_STATE_REASON_SERVICE_STOPPED:
return _("The VPN service stopped unexpectedly");
case NM_ACTIVE_CONNECTION_STATE_REASON_IP_CONFIG_INVALID:
return _("The VPN service returned invalid configuration");
case NM_ACTIVE_CONNECTION_STATE_REASON_CONNECT_TIMEOUT:
return _("The connection attempt timed out");
case NM_ACTIVE_CONNECTION_STATE_REASON_SERVICE_START_TIMEOUT:
return _("The VPN service did not start in time");
case NM_ACTIVE_CONNECTION_STATE_REASON_SERVICE_START_FAILED:
return _("The VPN service failed to start");
case NM_ACTIVE_CONNECTION_STATE_REASON_NO_SECRETS:
return _("No valid secrets");
case NM_ACTIVE_CONNECTION_STATE_REASON_LOGIN_FAILED:
return _("Invalid secrets");
case NM_ACTIVE_CONNECTION_STATE_REASON_CONNECTION_REMOVED:
return _("The connection was removed");
case NM_ACTIVE_CONNECTION_STATE_REASON_DEPENDENCY_FAILED:
return _("Master connection failed");
case NM_ACTIVE_CONNECTION_STATE_REASON_DEVICE_REALIZE_FAILED:
return _("Could not create a software link");
case NM_ACTIVE_CONNECTION_STATE_REASON_DEVICE_REMOVED:
return _("The device disappeared");
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
g_return_val_if_reached (_("Invalid reason"));
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
static void
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
check_activated (ActivateConnectionInfo *info)
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
{
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
NmCli *nmc = info->nmc;
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
NMDevice *device = info->device;
NMActiveConnection *active = info->active;
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
NMActiveConnectionState ac_state;
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
NMActiveConnectionStateReason ac_reason;
NMDeviceState dev_state;
NMDeviceStateReason dev_reason;
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
ac_state = nm_active_connection_get_state (active);
ac_reason = nm_active_connection_get_state_reason (active);
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
if (device) {
dev_state = nm_device_get_state (device);
dev_reason = nm_device_get_state_reason (device);
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
if (ac_state == NM_ACTIVE_CONNECTION_STATE_ACTIVATED) {
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output == NMC_PRINT_PRETTY)
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
nmc_terminal_erase_line ();
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Connection successfully activated (D-Bus active path: %s)\n"),
nm_object_get_path (NM_OBJECT (active)));
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
activate_connection_info_finish (info);
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
} else if (ac_state == NM_ACTIVE_CONNECTION_STATE_DEACTIVATED) {
if (device && ac_reason == NM_ACTIVE_CONNECTION_STATE_REASON_DEVICE_DISCONNECTED) {
if (dev_state == NM_DEVICE_STATE_FAILED || dev_state == NM_DEVICE_STATE_DISCONNECTED) {
g_string_printf (nmc->return_text, _("Error: Connection activation failed: %s"),
nmc_device_reason_to_string (dev_reason));
nmc->return_value = NMC_RESULT_ERROR_CON_ACTIVATION;
activate_connection_info_finish (info);
} else {
/* Just wait for the device to go failed. We'll get a better error message. */
return;
}
} else {
g_string_printf (nmc->return_text, _("Error: Connection activation failed: %s"),
active_connection_state_reason_to_string (ac_reason));
nmc->return_value = NMC_RESULT_ERROR_CON_ACTIVATION;
activate_connection_info_finish (info);
}
} else if (ac_state == NM_ACTIVE_CONNECTION_STATE_ACTIVATING) {
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
/* activating master connection does not automatically activate any slaves, so their
* active connection state will not progress beyond ACTIVATING state.
* Monitor the device instead. */
cli: Process secret agent request for a connection only if we know its path If we're activating the device without knowing the connection in advance, defer servicing the requests for the secrets until we set the path. [lkundrak@fedora20-2 ~]$ nmcli --ask c ifname wlan0 (process:18405): libnm-CRITICAL **: nm_object_get_path: assertion 'NM_IS_OBJECT (object)' failed Error: Connection activation failed: The device has no connections available. [lkundrak@fedora20-2 ~]$
2014-11-13 20:22:04 +01:00
if (nmc->secret_agent) {
NMRemoteConnection *connection = nm_active_connection_get_connection (active);
clients: merge nm_secret_agent_simple_set_connection_path() into nm_secret_agent_simple_enable() set_connection_path() is almost always called right before enable(), and it's unclear why it would be called anywhere else. So just merge the two methods.
2014-11-20 16:00:34 -06:00
nm_secret_agent_simple_enable (NM_SECRET_AGENT_SIMPLE (nmc->secret_agent),
nm_connection_get_path (NM_CONNECTION (connection)));
cli: Process secret agent request for a connection only if we know its path If we're activating the device without knowing the connection in advance, defer servicing the requests for the secrets until we set the path. [lkundrak@fedora20-2 ~]$ nmcli --ask c ifname wlan0 (process:18405): libnm-CRITICAL **: nm_object_get_path: assertion 'NM_IS_OBJECT (object)' failed Error: Connection activation failed: The device has no connections available. [lkundrak@fedora20-2 ~]$
2014-11-13 20:22:04 +01:00
}
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
if ( device
&& ( NM_IS_DEVICE_BOND (device)
|| NM_IS_DEVICE_TEAM (device)
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
|| NM_IS_DEVICE_BRIDGE (device))
&& dev_state >= NM_DEVICE_STATE_IP_CONFIG
&& dev_state <= NM_DEVICE_STATE_ACTIVATED) {
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output == NMC_PRINT_PRETTY)
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
nmc_terminal_erase_line ();
g_print (_("Connection successfully activated (master waiting for slaves) (D-Bus active path: %s)\n"),
nm_object_get_path (NM_OBJECT (active)));
activate_connection_info_finish (info);
cli: succeed in master bring-up when device state reaches IP_CONFIG (bgo #733641) When connecting a master connection, no slave devices will be activated automatically. The user is supposed to activate them individually. Hence nmcli should not wait for the connection to be fully activated because that is not going to happen (unless the user connects a slave connection from another terminal). Instead, for master connections behave differently and signal success once the master device reaches IP_CONFIG state. This revises behavior introduced by commit 47710f8211f178cddf5f84c1a50146d7476115a7. https://bugzilla.gnome.org/show_bug.cgi?id=733641 Signed-off-by: Thomas Haller <thaller@redhat.com> Signed-off-by: Jiří Klimeš <jklimes@redhat.com>
2014-07-28 14:01:23 +02:00
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
}
static void
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
device_state_cb (NMDevice *device, GParamSpec *pspec, ActivateConnectionInfo *info)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
check_activated (info);
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
static void
active_connection_state_cb (NMActiveConnection *active,
NMActiveConnectionState state,
NMActiveConnectionStateReason reason,
ActivateConnectionInfo *info)
{
check_activated (info);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: fix crash on connection delete/down timeout connection_op_timeout_cb() passed a wrong object type to timeout_cb(); fix this. Fixes: ac888de1510355e341fa8139692b918f2237fa53 https://bugzilla.redhat.com/show_bug.cgi?id=1355740
2016-07-18 18:50:24 +02:00
static void
set_nmc_error_timeout (NmCli *nmc)
{
g_string_printf (nmc->return_text, _("Error: Timeout expired (%d seconds)"), nmc->timeout);
nmc->return_value = NMC_RESULT_ERROR_TIMEOUT_EXPIRED;
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
static gboolean
cli: fix crash on connection delete/down timeout connection_op_timeout_cb() passed a wrong object type to timeout_cb(); fix this. Fixes: ac888de1510355e341fa8139692b918f2237fa53 https://bugzilla.redhat.com/show_bug.cgi?id=1355740
2016-07-18 18:50:24 +02:00
activate_connection_timeout_cb (gpointer user_data)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
ActivateConnectionInfo *info = user_data;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
/* Time expired -> exit nmcli */
cli: fix crash on connection delete/down timeout connection_op_timeout_cb() passed a wrong object type to timeout_cb(); fix this. Fixes: ac888de1510355e341fa8139692b918f2237fa53 https://bugzilla.redhat.com/show_bug.cgi?id=1355740
2016-07-18 18:50:24 +02:00
set_nmc_error_timeout (info->nmc);
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
activate_connection_info_finish (info);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
return FALSE;
}
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
static gboolean
progress_cb (gpointer user_data)
{
const char *str = (const char *) user_data;
nmc_terminal_show_progress (str);
return TRUE;
}
static gboolean
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
progress_active_connection_cb (gpointer user_data)
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
{
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
NMActiveConnection *active = user_data;
const char *str;
NMDevice *device;
NMActiveConnectionState ac_state;
const GPtrArray *ac_devs;
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
ac_state = nm_active_connection_get_state (active);
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
if (ac_state == NM_ACTIVE_CONNECTION_STATE_ACTIVATING) {
/* If the connection is activating, the device state
* is more interesting. */
ac_devs = nm_active_connection_get_devices (active);
device = ac_devs->len > 0 ? g_ptr_array_index (ac_devs, 0) : NULL;
} else {
device = NULL;
}
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
str = device
? nmc_device_state_to_string (nm_device_get_state (device))
: active_connection_state_to_string (ac_state);
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
nmc_terminal_show_progress (str);
return TRUE;
}
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
static void
activate_connection_info_finish (ActivateConnectionInfo *info)
{
cli: take a reference to the device when activating It can go away before the activation finishes: (process:378): GLib-GObject-WARNING **: instance with invalid (NULL) class pointer Thread 1 "lt-nmcli" received signal SIGTRAP, Trace/breakpoint trap. 0x00007ffff4fa919b in g_logv () from /lib64/libglib-2.0.so.0 (gdb) bt #0 0x00007ffff4fa919b in g_logv () at /lib64/libglib-2.0.so.0 #1 0x00007ffff4fa930f in g_log () at /lib64/libglib-2.0.so.0 #2 0x00007ffff529b3ea in g_type_check_instance () at /lib64/libgobject-2.0.so.0 #3 0x00007ffff528fa94 in g_signal_handlers_disconnect_matched () at /lib64/libgobject-2.0.so.0 #4 0x000000000041ae4e in activate_connection_info_finish (info=0x7b5490) at connections.c:2197 #5 0x00007ffff5279470 in g_cclosure_marshal_VOID__OBJECTv () at /lib64/libgobject-2.0.so.0 #6 0x00007ffff5276617 in _g_closure_invoke_va () at /lib64/libgobject-2.0.so.0 #7 0x00007ffff5290ea9 in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #8 0x00007ffff52914ff in g_signal_emit () at /lib64/libgobject-2.0.so.0 #9 0x00007ffff52763e8 in g_closure_invoke () at /lib64/libgobject-2.0.so.0 #10 0x00007ffff52884a2 in signal_emit_unlocked_R () at /lib64/libgobject-2.0.so.0 #11 0x00007ffff529112d in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #12 0x00007ffff52919ab in g_signal_emit_by_name () at /lib64/libgobject-2.0.so.0 #13 0x00007ffff78fc9b1 in deferred_notify_cb (data=<optimized out>) at nm-object.c:243 #14 0x00007ffff4fa2703 in g_main_context_dispatch () at /lib64/libglib-2.0.so.0 #15 0x00007ffff4fa2ab0 in g_main_context_iterate.isra () at /lib64/libglib-2.0.so.0 #16 0x00007ffff4fa2dd2 in g_main_loop_run () at /lib64/libglib-2.0.so.0 #17 0x00000000004146e3 in main (argc=<optimized out>, argv=<optimized out>) at nmcli.c:618 (gdb) print info->device
2016-07-28 11:21:29 +02:00
if (info->device) {
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
g_signal_handlers_disconnect_by_func (info->device, G_CALLBACK (device_state_cb), info);
cli: take a reference to the device when activating It can go away before the activation finishes: (process:378): GLib-GObject-WARNING **: instance with invalid (NULL) class pointer Thread 1 "lt-nmcli" received signal SIGTRAP, Trace/breakpoint trap. 0x00007ffff4fa919b in g_logv () from /lib64/libglib-2.0.so.0 (gdb) bt #0 0x00007ffff4fa919b in g_logv () at /lib64/libglib-2.0.so.0 #1 0x00007ffff4fa930f in g_log () at /lib64/libglib-2.0.so.0 #2 0x00007ffff529b3ea in g_type_check_instance () at /lib64/libgobject-2.0.so.0 #3 0x00007ffff528fa94 in g_signal_handlers_disconnect_matched () at /lib64/libgobject-2.0.so.0 #4 0x000000000041ae4e in activate_connection_info_finish (info=0x7b5490) at connections.c:2197 #5 0x00007ffff5279470 in g_cclosure_marshal_VOID__OBJECTv () at /lib64/libgobject-2.0.so.0 #6 0x00007ffff5276617 in _g_closure_invoke_va () at /lib64/libgobject-2.0.so.0 #7 0x00007ffff5290ea9 in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #8 0x00007ffff52914ff in g_signal_emit () at /lib64/libgobject-2.0.so.0 #9 0x00007ffff52763e8 in g_closure_invoke () at /lib64/libgobject-2.0.so.0 #10 0x00007ffff52884a2 in signal_emit_unlocked_R () at /lib64/libgobject-2.0.so.0 #11 0x00007ffff529112d in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #12 0x00007ffff52919ab in g_signal_emit_by_name () at /lib64/libgobject-2.0.so.0 #13 0x00007ffff78fc9b1 in deferred_notify_cb (data=<optimized out>) at nm-object.c:243 #14 0x00007ffff4fa2703 in g_main_context_dispatch () at /lib64/libglib-2.0.so.0 #15 0x00007ffff4fa2ab0 in g_main_context_iterate.isra () at /lib64/libglib-2.0.so.0 #16 0x00007ffff4fa2dd2 in g_main_loop_run () at /lib64/libglib-2.0.so.0 #17 0x00000000004146e3 in main (argc=<optimized out>, argv=<optimized out>) at nmcli.c:618 (gdb) print info->device
2016-07-28 11:21:29 +02:00
g_object_unref (info->device);
}
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
if (info->active) {
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
g_signal_handlers_disconnect_by_func (info->active, G_CALLBACK (active_connection_state_cb), info);
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
g_object_unref (info->active);
}
g_free (info);
quit ();
}
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
static void
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
activate_connection_cb (GObject *client, GAsyncResult *result, gpointer user_data)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
ActivateConnectionInfo *info = (ActivateConnectionInfo *) user_data;
NmCli *nmc = info->nmc;
NMDevice *device = info->device;
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
NMActiveConnection *active;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
NMActiveConnectionState state;
cli: fix erroneous messages and asserts for activation bond/bridge connections error messages in bond_bridge_slaves_check(): ** (process:11921): CRITICAL **: nm_device_bond_get_slaves: assertion `NM_IS_DEVICE_BOND (device)' failed ** (process:11921): CRITICAL **: nm_device_get_iface: assertion `NM_IS_DEVICE (device)' failed
2013-06-26 13:38:58 +02:00
const GPtrArray *ac_devs;
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
GError *error = NULL;
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
info->active = active = nm_client_activate_connection_finish (NM_CLIENT (client), result, &error);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: update due to changes in libnm-glib's NMClient nm_client_activate_connection() now uses NMConnection instead of D-Bus connection path. Also, callback uses NMActiveConnection instead of the path. We can also remove the workaround waiting for "active-connections" change, because the activation callback is now called *after* the NMActiveConnection object is available in the active connections list.
2011-03-16 13:29:41 +01:00
if (error) {
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
g_string_printf (nmc->return_text, _("Error: Connection activation failed: %s"),
error->message);
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
g_error_free (error);
cli: update due to changes in libnm-glib's NMClient nm_client_activate_connection() now uses NMConnection instead of D-Bus connection path. Also, callback uses NMActiveConnection instead of the path. We can also remove the workaround waiting for "active-connections" change, because the activation callback is now called *after* the NMActiveConnection object is available in the active connections list.
2011-03-16 13:29:41 +01:00
nmc->return_value = NMC_RESULT_ERROR_CON_ACTIVATION;
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
activate_connection_info_finish (info);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
} else {
state = nm_active_connection_get_state (active);
cli: fix erroneous messages and asserts for activation bond/bridge connections error messages in bond_bridge_slaves_check(): ** (process:11921): CRITICAL **: nm_device_bond_get_slaves: assertion `NM_IS_DEVICE_BOND (device)' failed ** (process:11921): CRITICAL **: nm_device_get_iface: assertion `NM_IS_DEVICE (device)' failed
2013-06-26 13:38:58 +02:00
if (!device) {
/* device could be NULL for virtual devices. Fill it here. */
ac_devs = nm_active_connection_get_devices (active);
cli: take a reference to the device when activating It can go away before the activation finishes: (process:378): GLib-GObject-WARNING **: instance with invalid (NULL) class pointer Thread 1 "lt-nmcli" received signal SIGTRAP, Trace/breakpoint trap. 0x00007ffff4fa919b in g_logv () from /lib64/libglib-2.0.so.0 (gdb) bt #0 0x00007ffff4fa919b in g_logv () at /lib64/libglib-2.0.so.0 #1 0x00007ffff4fa930f in g_log () at /lib64/libglib-2.0.so.0 #2 0x00007ffff529b3ea in g_type_check_instance () at /lib64/libgobject-2.0.so.0 #3 0x00007ffff528fa94 in g_signal_handlers_disconnect_matched () at /lib64/libgobject-2.0.so.0 #4 0x000000000041ae4e in activate_connection_info_finish (info=0x7b5490) at connections.c:2197 #5 0x00007ffff5279470 in g_cclosure_marshal_VOID__OBJECTv () at /lib64/libgobject-2.0.so.0 #6 0x00007ffff5276617 in _g_closure_invoke_va () at /lib64/libgobject-2.0.so.0 #7 0x00007ffff5290ea9 in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #8 0x00007ffff52914ff in g_signal_emit () at /lib64/libgobject-2.0.so.0 #9 0x00007ffff52763e8 in g_closure_invoke () at /lib64/libgobject-2.0.so.0 #10 0x00007ffff52884a2 in signal_emit_unlocked_R () at /lib64/libgobject-2.0.so.0 #11 0x00007ffff529112d in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #12 0x00007ffff52919ab in g_signal_emit_by_name () at /lib64/libgobject-2.0.so.0 #13 0x00007ffff78fc9b1 in deferred_notify_cb (data=<optimized out>) at nm-object.c:243 #14 0x00007ffff4fa2703 in g_main_context_dispatch () at /lib64/libglib-2.0.so.0 #15 0x00007ffff4fa2ab0 in g_main_context_iterate.isra () at /lib64/libglib-2.0.so.0 #16 0x00007ffff4fa2dd2 in g_main_loop_run () at /lib64/libglib-2.0.so.0 #17 0x00000000004146e3 in main (argc=<optimized out>, argv=<optimized out>) at nmcli.c:618 (gdb) print info->device
2016-07-28 11:21:29 +02:00
device = ac_devs->len > 0 ? g_ptr_array_index (ac_devs, 0) : NULL;
if (device)
info->device = g_object_ref (device);
cli: fix erroneous messages and asserts for activation bond/bridge connections error messages in bond_bridge_slaves_check(): ** (process:11921): CRITICAL **: nm_device_bond_get_slaves: assertion `NM_IS_DEVICE_BOND (device)' failed ** (process:11921): CRITICAL **: nm_device_get_iface: assertion `NM_IS_DEVICE (device)' failed
2013-06-26 13:38:58 +02:00
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: update due to changes in libnm-glib's NMClient nm_client_activate_connection() now uses NMConnection instead of D-Bus connection path. Also, callback uses NMActiveConnection instead of the path. We can also remove the workaround waiting for "active-connections" change, because the activation callback is now called *after* the NMActiveConnection object is available in the active connections list.
2011-03-16 13:29:41 +01:00
if (nmc->nowait_flag || state == NM_ACTIVE_CONNECTION_STATE_ACTIVATED) {
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
/* User doesn't want to wait or already activated */
cli: always print success message (not only in --pretty mode) (rh #1006444) The commands performing actions without an output only printed a successfull message for --pretty mode. Some users don't like this silent operation. This commit makes nmcli print operation success unconditionally. I think that doesn't hurt; scripts/users can ignore the message if not interested. https://bugzilla.redhat.com/show_bug.cgi?id=1006444
2013-09-19 15:15:10 +02:00
if (state == NM_ACTIVE_CONNECTION_STATE_ACTIVATED) {
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output == NMC_PRINT_PRETTY)
cli: always print success message (not only in --pretty mode) (rh #1006444) The commands performing actions without an output only printed a successfull message for --pretty mode. Some users don't like this silent operation. This commit makes nmcli print operation success unconditionally. I think that doesn't hurt; scripts/users can ignore the message if not interested. https://bugzilla.redhat.com/show_bug.cgi?id=1006444
2013-09-19 15:15:10 +02:00
nmc_terminal_erase_line ();
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Connection successfully activated (D-Bus active path: %s)\n"),
nm_object_get_path (NM_OBJECT (active)));
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
}
cli/connections: fail the activation when the active connection disappears Relying solemnly on device or active connection state is unreliable.
2016-07-04 17:56:20 +02:00
activate_connection_info_finish (info);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
} else {
nmcli: avoid calling twice "check_activated()" on "nmcli connection up" This happens when the connection is in "activating" state and the connection is a master one waiting for slaves: "check_activated()" is called by the active_connection_state_cb() and device_state_cb() callbacks. If the device has already moved to a state >= NM_DEVICE_STATE_IP_CONFIG, the call to active_connection_state_cb() will end calling activate_connection_info_finish(), freeing the "info" object. The subsequent call to device_state_cb() will result in accessing the freed "info". Just call check_activated() once after registering the active_connection_state() and device_state_cb() callbacks.
2017-04-05 10:48:16 +02:00
/* Monitor the active connection and device (if available) states */
g_signal_connect (active, "state-changed", G_CALLBACK (active_connection_state_cb), info);
if (device)
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
g_signal_connect (device, "notify::" NM_DEVICE_STATE, G_CALLBACK (device_state_cb), info);
nmcli: avoid calling twice "check_activated()" on "nmcli connection up" This happens when the connection is in "activating" state and the connection is a master one waiting for slaves: "check_activated()" is called by the active_connection_state_cb() and device_state_cb() callbacks. If the device has already moved to a state >= NM_DEVICE_STATE_IP_CONFIG, the call to active_connection_state_cb() will end calling activate_connection_info_finish(), freeing the "info" object. The subsequent call to device_state_cb() will result in accessing the freed "info". Just call check_activated() once after registering the active_connection_state() and device_state_cb() callbacks.
2017-04-05 10:48:16 +02:00
/* Both active_connection_state_cb () and device_state_cb () will just
* call check_activated (info). So, just call it once directly after
* connecting on both the signals of the objects and skip the call to
* the callbacks.
*/
check_activated (info);
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
/* Start progress indication showing VPN states */
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output == NMC_PRINT_PRETTY) {
cli/connections: decide about activation success and failure in single place Track both device and active connections at the same time and decide whether activation finished (either successfully or not) in a single place, check_activated(). This makes the already too complex code path a bit more straightforward and also makes it possible to be more reasonable about the diagnostics. Now that the active connection signals the reason, we include it; but if the failure is due to the device disconnection while we're activating, include a device reason instead, since it's often more useful. Like: Before: Error: Connection activation failed. Without considering the device: Error: Connection activation failed: the base network connection was interrupted After: Error: Connection activation failed: The Wi-Fi network could not be found
2017-03-06 11:50:06 +01:00
if (progress_id)
g_source_remove (progress_id);
progress_id = g_timeout_add (120, progress_active_connection_cb, active);
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
/* Start timer not to loop forever when signals are not emitted */
cli: fix crash on connection delete/down timeout connection_op_timeout_cb() passed a wrong object type to timeout_cb(); fix this. Fixes: ac888de1510355e341fa8139692b918f2237fa53 https://bugzilla.redhat.com/show_bug.cgi?id=1355740
2016-07-18 18:50:24 +02:00
g_timeout_add_seconds (nmc->timeout, activate_connection_timeout_cb, info);
cli: fix erroneous messages and asserts for activation bond/bridge connections error messages in bond_bridge_slaves_check(): ** (process:11921): CRITICAL **: nm_device_bond_get_slaves: assertion `NM_IS_DEVICE_BOND (device)' failed ** (process:11921): CRITICAL **: nm_device_get_iface: assertion `NM_IS_DEVICE (device)' failed
2013-06-26 13:38:58 +02:00
}
cli: 'con up' - give bond/bridge slaves 10s to appear while activating master If there are no slaves after the timeout, exit nmcli with an error, assuming that NM is waiting for slaves to finish activation: "(bondx) "IPv4 config waiting until slaves are ready" "(bondx) "IPv6 config waiting until slaves are ready"
2013-04-29 12:43:13 +02:00
}
}
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
/**
* parse_passwords:
* @passwd_file: file with passwords to parse
* @error: location to store error, or %NULL
*
* Parse passwords given in @passwd_file and insert them into a hash table.
* Example of @passwd_file contents:
* wifi.psk:tajne heslo
* 802-1x.password:krakonos
* 802-11-wireless-security:leap-password:my leap password
*
* Returns: hash table with parsed passwords, or %NULL on an error
*/
static GHashTable *
parse_passwords (const char *passwd_file, GError **error)
{
GHashTable *pwds_hash;
char *contents = NULL;
gsize len = 0;
GError *local_err = NULL;
char **lines, **iter;
char *pwd_spec, *pwd, *prop;
const char *setting;
pwds_hash = g_hash_table_new_full (g_str_hash, g_str_equal, g_free, g_free);
if (!passwd_file)
return pwds_hash;
/* Read the passwords file */
if (!g_file_get_contents (passwd_file, &contents, &len, &local_err)) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("failed to read passwd-file '%s': %s"),
passwd_file, local_err->message);
g_error_free (local_err);
g_hash_table_destroy (pwds_hash);
return NULL;
}
lines = nmc_strsplit_set (contents, "\r\n", -1);
for (iter = lines; *iter; iter++) {
pwd = strchr (*iter, ':');
if (!pwd) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("missing colon in 'password' entry '%s'"), *iter);
goto failure;
}
*(pwd++) = '\0';
prop = strchr (*iter, '.');
if (!prop) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("missing dot in 'password' entry '%s'"), *iter);
goto failure;
}
*(prop++) = '\0';
setting = *iter;
while (g_ascii_isspace (*setting))
setting++;
/* Accept wifi-sec or wifi instead of cumbersome '802-11-wireless-security' */
if (!strcmp (setting, "wifi-sec") || !strcmp (setting, "wifi"))
setting = NM_SETTING_WIRELESS_SECURITY_SETTING_NAME;
if (nm_setting_lookup_type (setting) == G_TYPE_INVALID) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("invalid setting name in 'password' entry '%s'"), setting);
goto failure;
}
pwd_spec = g_strdup_printf ("%s.%s", setting, prop);
g_hash_table_insert (pwds_hash, pwd_spec, g_strdup (pwd));
}
g_strfreev (lines);
g_free (contents);
return pwds_hash;
failure:
g_strfreev (lines);
g_free (contents);
g_hash_table_destroy (pwds_hash);
return NULL;
}
cli: use secret agent for getting passwords from user --ask option has to be used, so that nmcli can be interactive.
2014-10-06 15:25:53 +02:00
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
static gboolean
nmc_activate_connection (NmCli *nmc,
NMConnection *connection,
const char *ifname,
const char *ap,
const char *nsp,
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
const char *pwds,
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
GAsyncReadyCallback callback,
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
GError **error)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
ActivateConnectionInfo *info;
clients: only handle secret requests for connection being explicitly activated When a connection is being activated, nmcli could ask for secrets for another connection, which might confuse users. We check the request now and only ask for secrets of connection being activated. Test case: $ nmcli con up my-ethernet0 Passwords or encryption keys are required to access the wireless network 'Red Hat'. Warning: password for '802-1x.identity' not given in 'passwd-file' and nmcli cannot ask without '--ask' option.
2014-11-11 13:15:19 +01:00
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
GHashTable *pwds_hash;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
NMDevice *device = NULL;
const char *spec_object = NULL;
gboolean device_found;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
GError *local = NULL;
g_return_val_if_fail (nmc != NULL, FALSE);
g_return_val_if_fail (error == NULL || *error == NULL, FALSE);
cli: don't look up a device for activation request unless we have to Let the server decide which device to use if the user didn't explicitly specify the interface, wireless access point or a wimax nsp. The server will just reuse the device for an already active connection or potentially do a better guess. https://bugzilla.gnome.org/show_bug.cgi?id=730492
2015-02-01 09:09:37 +01:00
if (connection && (ifname || ap || nsp)) {
cli: add support for 'nmcli con up ifname XXX' Passes a NULL connection to nm_client_activate_connection() allowing NetworkManager to pick the best available connection for the interface.
2013-09-06 10:05:47 -05:00
device_found = find_device_for_connection (nmc, connection, ifname, ap, nsp, &device, &spec_object, &local);
nmcli: fix connecting VLANs without an explicit interface-name (rh #1034908) nm_connection_get_virtual_iface_name() doesn't work when determining virtual connections, because for VLANs it can return NULL. See also commit e1e4740648d3ee522c8a80d1af6282afce94f53d. https://bugzilla.redhat.com/show_bug.cgi?id=1034908
2013-11-28 16:48:06 +01:00
cli: add support for 'nmcli con up ifname XXX' Passes a NULL connection to nm_client_activate_connection() allowing NetworkManager to pick the best available connection for the interface.
2013-09-06 10:05:47 -05:00
/* Virtual connection may not have their interfaces created yet */
libnm-core: add nm_connection_is_virtual() Add a method to determine if a connection applies to a virtual device. Perhaps eventually the logic should be spread across the NMSetting classes, but for now it's better off having it in NMConnection than once in NMManager and once in nmcli.
2014-08-04 16:27:43 -04:00
if (!device_found && !nm_connection_is_virtual (connection)) {
cli: add support for 'nmcli con up ifname XXX' Passes a NULL connection to nm_client_activate_connection() allowing NetworkManager to pick the best available connection for the interface.
2013-09-06 10:05:47 -05:00
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_CON_ACTIVATION,
all: clean-up usage of GError Functions that take a GError** MUST fill it in on error. There is no need to check whether error is NULL if the function it was passed to had a failing return value. Likewise, a proper GError must have a non-NULL message, so there's no need to double-check that either. Based-on-patch-by: Dan Winship <danw@gnome.org>
2016-02-28 16:25:36 +01:00
"%s", local->message);
cli: add support for 'nmcli con up ifname XXX' Passes a NULL connection to nm_client_activate_connection() allowing NetworkManager to pick the best available connection for the interface.
2013-09-06 10:05:47 -05:00
g_clear_error (&local);
return FALSE;
}
cli: fix an activation bug Add a missing g_clear_error() that, for circuitous reasons, made trying to activate a virtual connection fail silently.
2014-11-07 13:56:49 -05:00
g_clear_error (&local);
cli: add support for 'nmcli con up ifname XXX' Passes a NULL connection to nm_client_activate_connection() allowing NetworkManager to pick the best available connection for the interface.
2013-09-06 10:05:47 -05:00
} else if (ifname) {
device = nm_client_get_device_by_iface (nmc->client, ifname);
if (!device) {
cli: add NMC_RESULT_ERROR_NOT_FOUND and use as appropriate This will allow callers to distinguish "no such connection" from "connection failed to activate", etc. https://bugzilla.redhat.com/show_bug.cgi?id=1022256
2013-11-21 09:19:18 -05:00
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_NOT_FOUND,
cli: add support for 'nmcli con up ifname XXX' Passes a NULL connection to nm_client_activate_connection() allowing NetworkManager to pick the best available connection for the interface.
2013-09-06 10:05:47 -05:00
_("unknown device '%s'."), ifname);
return FALSE;
}
cli: don't look up a device for activation request unless we have to Let the server decide which device to use if the user didn't explicitly specify the interface, wireless access point or a wimax nsp. The server will just reuse the device for an already active connection or potentially do a better guess. https://bugzilla.gnome.org/show_bug.cgi?id=730492
2015-02-01 09:09:37 +01:00
} else if (!connection) {
cli: add NMC_RESULT_ERROR_NOT_FOUND and use as appropriate This will allow callers to distinguish "no such connection" from "connection failed to activate", etc. https://bugzilla.redhat.com/show_bug.cgi?id=1022256
2013-11-21 09:19:18 -05:00
g_set_error_literal (error, NMCLI_ERROR, NMC_RESULT_ERROR_NOT_FOUND,
cli: fix 'nmcli con up bad_name' or 'nmcli con up' ** (process:25157): CRITICAL **: nmc_activate_connection: assertion `NM_IS_CONNECTION (connection) || ifname' failed Error: unknown error. Regression caused by 42c7ea85a17a7e7459942f356399e36308011c86.
2013-11-07 18:01:16 +01:00
_("neither a valid connection nor device given"));
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
return FALSE;
}
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
/* Parse passwords given in passwords file */
pwds_hash = parse_passwords (pwds, &local);
if (local) {
g_propagate_error (error, local);
return FALSE;
}
if (nmc->pwds_hash)
g_hash_table_destroy (nmc->pwds_hash);
nmc->pwds_hash = pwds_hash;
cli: use secret agent for getting passwords from user --ask option has to be used, so that nmcli can be interactive.
2014-10-06 15:25:53 +02:00
/* Create secret agent */
cli: Process secret agent request for a connection only if we know its path If we're activating the device without knowing the connection in advance, defer servicing the requests for the secrets until we set the path. [lkundrak@fedora20-2 ~]$ nmcli --ask c ifname wlan0 (process:18405): libnm-CRITICAL **: nm_object_get_path: assertion 'NM_IS_OBJECT (object)' failed Error: Connection activation failed: The device has no connections available. [lkundrak@fedora20-2 ~]$
2014-11-13 20:22:04 +01:00
nmc->secret_agent = nm_secret_agent_simple_new ("nmcli-connect");
if (nmc->secret_agent) {
clients: add define for NMSecretAgentSimple signal name
2016-09-09 17:33:09 +02:00
g_signal_connect (nmc->secret_agent,
NM_SECRET_AGENT_SIMPLE_REQUEST_SECRETS,
G_CALLBACK (nmc_secrets_requested),
nmc);
cli: Process secret agent request for a connection only if we know its path If we're activating the device without knowing the connection in advance, defer servicing the requests for the secrets until we set the path. [lkundrak@fedora20-2 ~]$ nmcli --ask c ifname wlan0 (process:18405): libnm-CRITICAL **: nm_object_get_path: assertion 'NM_IS_OBJECT (object)' failed Error: Connection activation failed: The device has no connections available. [lkundrak@fedora20-2 ~]$
2014-11-13 20:22:04 +01:00
if (connection) {
clients: merge nm_secret_agent_simple_set_connection_path() into nm_secret_agent_simple_enable() set_connection_path() is almost always called right before enable(), and it's unclear why it would be called anywhere else. So just merge the two methods.
2014-11-20 16:00:34 -06:00
nm_secret_agent_simple_enable (NM_SECRET_AGENT_SIMPLE (nmc->secret_agent),
nm_object_get_path (NM_OBJECT (connection)));
cli: Process secret agent request for a connection only if we know its path If we're activating the device without knowing the connection in advance, defer servicing the requests for the secrets until we set the path. [lkundrak@fedora20-2 ~]$ nmcli --ask c ifname wlan0 (process:18405): libnm-CRITICAL **: nm_object_get_path: assertion 'NM_IS_OBJECT (object)' failed Error: Connection activation failed: The device has no connections available. [lkundrak@fedora20-2 ~]$
2014-11-13 20:22:04 +01:00
}
}
cli: use secret agent for getting passwords from user --ask option has to be used, so that nmcli can be interactive.
2014-10-06 15:25:53 +02:00
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
info = g_malloc0 (sizeof (ActivateConnectionInfo));
info->nmc = nmc;
cli: take a reference to the device when activating It can go away before the activation finishes: (process:378): GLib-GObject-WARNING **: instance with invalid (NULL) class pointer Thread 1 "lt-nmcli" received signal SIGTRAP, Trace/breakpoint trap. 0x00007ffff4fa919b in g_logv () from /lib64/libglib-2.0.so.0 (gdb) bt #0 0x00007ffff4fa919b in g_logv () at /lib64/libglib-2.0.so.0 #1 0x00007ffff4fa930f in g_log () at /lib64/libglib-2.0.so.0 #2 0x00007ffff529b3ea in g_type_check_instance () at /lib64/libgobject-2.0.so.0 #3 0x00007ffff528fa94 in g_signal_handlers_disconnect_matched () at /lib64/libgobject-2.0.so.0 #4 0x000000000041ae4e in activate_connection_info_finish (info=0x7b5490) at connections.c:2197 #5 0x00007ffff5279470 in g_cclosure_marshal_VOID__OBJECTv () at /lib64/libgobject-2.0.so.0 #6 0x00007ffff5276617 in _g_closure_invoke_va () at /lib64/libgobject-2.0.so.0 #7 0x00007ffff5290ea9 in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #8 0x00007ffff52914ff in g_signal_emit () at /lib64/libgobject-2.0.so.0 #9 0x00007ffff52763e8 in g_closure_invoke () at /lib64/libgobject-2.0.so.0 #10 0x00007ffff52884a2 in signal_emit_unlocked_R () at /lib64/libgobject-2.0.so.0 #11 0x00007ffff529112d in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0 #12 0x00007ffff52919ab in g_signal_emit_by_name () at /lib64/libgobject-2.0.so.0 #13 0x00007ffff78fc9b1 in deferred_notify_cb (data=<optimized out>) at nm-object.c:243 #14 0x00007ffff4fa2703 in g_main_context_dispatch () at /lib64/libglib-2.0.so.0 #15 0x00007ffff4fa2ab0 in g_main_context_iterate.isra () at /lib64/libglib-2.0.so.0 #16 0x00007ffff4fa2dd2 in g_main_loop_run () at /lib64/libglib-2.0.so.0 #17 0x00000000004146e3 in main (argc=<optimized out>, argv=<optimized out>) at nmcli.c:618 (gdb) print info->device
2016-07-28 11:21:29 +02:00
if (device)
info->device = g_object_ref (device);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
nm_client_activate_connection_async (nmc->client,
connection,
device,
spec_object,
NULL,
callback,
info);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
return TRUE;
}
static NMCResultCode
do_connection_up (NmCli *nmc, int argc, char **argv)
{
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
NMConnection *connection = NULL;
cli: rename 'iface' --> 'ifname' in all commands
2013-04-18 14:54:01 +02:00
const char *ifname = NULL;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
const char *ap = NULL;
cli: add support for WiMAX devices and connections
2011-01-06 17:01:55 -06:00
const char *nsp = NULL;
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
const char *pwds = NULL;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
gs_free_error GError *error = NULL;
char **arg_arr = NULL;
int arg_num;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
char ***argv_ptr;
int *argc_ptr;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: remove '--nowait' and '--timeout' options and use global '--wait' instead '--wait 0' corresponds to '--nowait' '--wait n' for n > 0 corresponds to '--timeout n'
2013-05-14 12:43:43 +02:00
/*
* Set default timeout for connection activation.
* Activation can take quite a long time, use 90 seconds.
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
*/
cli: remove '--nowait' and '--timeout' options and use global '--wait' instead '--wait 0' corresponds to '--nowait' '--wait n' for n > 0 corresponds to '--timeout n'
2013-05-14 12:43:43 +02:00
if (nmc->timeout == -1)
nmc->timeout = 90;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
argv_ptr = &argv;
argc_ptr = &argc;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
if (argc == 0 && nmc->ask) {
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
char *line;
/* nmc_do_cmd() should not call this with argc=0. */
g_assert (!nmc->complete);
line = nmc_readline ("%s: ", PROMPT_CONNECTION);
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
nmc_string_to_arg_array (line, NULL, TRUE, &arg_arr, &arg_num);
g_free (line);
argv_ptr = &arg_arr;
argc_ptr = &arg_num;
cli: make id|uuid optional for 'nmcli connection up' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection up [id|uuid|path] <ID> Examples: nmcli connection up "My Home Wi-Fi" nmcli connection up id "My Home Wi-Fi" nmcli connection up path /org/freedesktop/NetworkManager/Settings/18 nmcli connection up path 18 Note: In order to be able to identify connections with just index number, the 'path' keyword has to be provided.
2012-11-19 12:31:33 +01:00
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
if (argc > 0 && strcmp (*argv, "ifname") != 0) {
connection = get_connection (nmc, argc_ptr, argv_ptr, NULL, &error);
cli: Abort when given name of a non-existent connection for nmcli up Even when ifname is present: nmcli c up name Nonexistent ifname wlan0
2014-11-13 20:21:02 +01:00
if (!connection) {
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_string_printf (nmc->return_text, _("Error: %s."), error->message);
return error->code;
cli: Abort when given name of a non-existent connection for nmcli up Even when ifname is present: nmcli c up name Nonexistent ifname wlan0
2014-11-13 20:21:02 +01:00
}
}
cli: make id|uuid optional for 'nmcli connection up' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection up [id|uuid|path] <ID> Examples: nmcli connection up "My Home Wi-Fi" nmcli connection up id "My Home Wi-Fi" nmcli connection up path /org/freedesktop/NetworkManager/Settings/18 nmcli connection up path 18 Note: In order to be able to identify connections with just index number, the 'path' keyword has to be provided.
2012-11-19 12:31:33 +01:00
while (argc > 0) {
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
if (argc == 1 && nmc->complete)
nmc_complete_strings (*argv, "ifname", "ap", "passwd-file", NULL);
cli: rename 'iface' --> 'ifname' in all commands
2013-04-18 14:54:01 +02:00
if (strcmp (*argv, "ifname") == 0) {
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
cli: allow next_arg() to move arguments past the last one
2012-11-16 22:43:29 +01:00
g_string_printf (nmc->return_text, _("Error: %s argument is missing."), *(argv-1));
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
return NMC_RESULT_ERROR_USER_INPUT;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: rename 'iface' --> 'ifname' in all commands
2013-04-18 14:54:01 +02:00
ifname = *argv;
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
if (argc == 1 && nmc->complete)
nmc_complete_device (nmc->client, ifname, ap != NULL);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
else if (strcmp (*argv, "ap") == 0) {
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
cli: allow next_arg() to move arguments past the last one
2012-11-16 22:43:29 +01:00
g_string_printf (nmc->return_text, _("Error: %s argument is missing."), *(argv-1));
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
return NMC_RESULT_ERROR_USER_INPUT;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
ap = *argv;
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
if (argc == 1 && nmc->complete)
nmc_complete_bssid (nmc->client, ifname, ap);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
else if (strcmp (*argv, "passwd-file") == 0) {
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
g_string_printf (nmc->return_text, _("Error: %s argument is missing."), *(argv-1));
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
return NMC_RESULT_ERROR_USER_INPUT;
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
}
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
if (argc == 1 && nmc->complete)
nmc->return_value = NMC_RESULT_COMPLETE_FILE;
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
pwds = *argv;
}
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
else if (!nmc->complete) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Unknown parameter: %s\n"), *argv);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli/connections: add completion to up subcommand
2016-07-12 00:21:48 +02:00
if (nmc->complete)
return nmc->return_value;
cli: update due to changes in libnm-glib's NMClient nm_client_activate_connection() now uses NMConnection instead of D-Bus connection path. Also, callback uses NMActiveConnection instead of the path. We can also remove the workaround waiting for "active-connections" change, because the activation callback is now called *after* the NMActiveConnection object is available in the active connections list.
2011-03-16 13:29:41 +01:00
/* Use nowait_flag instead of should_wait because exiting has to be postponed till
* active_connection_state_cb() is called. That gives NM time to check our permissions
* and we can follow activation progress.
*/
cli: remove '--nowait' and '--timeout' options and use global '--wait' instead '--wait 0' corresponds to '--nowait' '--wait n' for n > 0 corresponds to '--timeout n'
2013-05-14 12:43:43 +02:00
nmc->nowait_flag = (nmc->timeout == 0);
cli: turn should_wait into a semaphore Count the exit blockers. We'll want to terminate the device monitor in case all monitored devices vanish.
2015-03-27 12:39:17 +01:00
nmc->should_wait++;
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
if (!nmc_activate_connection (nmc, connection, ifname, ap, nsp, pwds, activate_connection_cb, &error)) {
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
g_string_printf (nmc->return_text, _("Error: %s."),
all: clean-up usage of GError Functions that take a GError** MUST fill it in on error. There is no need to check whether error is NULL if the function it was passed to had a failing return value. Likewise, a proper GError must have a non-NULL message, so there's no need to double-check that either. Based-on-patch-by: Dan Winship <danw@gnome.org>
2016-02-28 16:25:36 +01:00
error->message);
cli: turn should_wait into a semaphore Count the exit blockers. We'll want to terminate the device monitor in case all monitored devices vanish.
2015-03-27 12:39:17 +01:00
nmc->should_wait--;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
return error->code;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
}
cli: show progress in --pretty mode for 'nmcli con up' We need to use a workaround for VPN connection state. There is a bug somewhere causing that vpn-state-changed signal is not issued on active VPN connection. Debugging showed that D-Bus PropertiesChanged and VpnStateChanged signals are not processed properly. That's why libnm-glib's VPN connection object is not updated. The issue may lie in libnm-glib or even in dbus-glib. I wasn't able to get VpnStateChanged signal via dbus-glib (dbus_g_proxy_add_signal() and dbus_g_proxy_connect_signal()), however received the signal successfully using GDBus. Until the issue is analyzed and fixed, we need to ask for VPN state of active VPN connection every second in order to show progress and exit correctly.
2012-05-09 12:35:27 +02:00
/* Start progress indication */
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output == NMC_PRINT_PRETTY)
cli: correct printing success indication on 'nmcli con up'
2013-07-03 12:43:36 +02:00
progress_id = g_timeout_add (120, progress_cb, _("preparing"));
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
return nmc->return_value;
}
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
typedef struct {
NmCli *nmc;
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
GSList *queue;
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
guint timeout_id;
cli: avoid use-after free on connection deletion If the connection spontaneously disappears (perhaps along with the whole daemon on crash) while we're deleting it, then the removal callback would free up the context structure the delete operation is using. Let's cancel the in-flight delete operations so that they won't touch the structure after it's gone.
2017-01-26 15:59:01 +01:00
GCancellable *cancellable;
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
} ConnectionCbInfo;
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
static void connection_cb_info_finish (ConnectionCbInfo *info,
gpointer connection);
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
static void
connection_removed_cb (NMClient *client, NMConnection *connection, ConnectionCbInfo *info)
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
{
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
if (!g_slist_find (info->queue, connection))
return;
g_print (_("Connection '%s' (%s) successfully deleted.\n"),
nm_connection_get_id (connection),
nm_connection_get_uuid (connection));
connection_cb_info_finish (info, connection);
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
}
static void
down_active_connection_state_cb (NMActiveConnection *active,
GParamSpec *pspec,
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
ConnectionCbInfo *info)
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
{
if (nm_active_connection_get_state (active) < NM_ACTIVE_CONNECTION_STATE_DEACTIVATED)
return;
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (info->nmc->nmc_config.print_output == NMC_PRINT_PRETTY)
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
nmc_terminal_erase_line ();
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
g_print (_("Connection '%s' successfully deactivated (D-Bus active path: %s)\n"),
nm_active_connection_get_id (active), nm_object_get_path (NM_OBJECT (active)));
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
g_signal_handlers_disconnect_by_func (G_OBJECT (active),
down_active_connection_state_cb,
info);
connection_cb_info_finish (info, active);
}
static gboolean
connection_op_timeout_cb (gpointer user_data)
{
ConnectionCbInfo *info = user_data;
cli: fix crash on connection delete/down timeout connection_op_timeout_cb() passed a wrong object type to timeout_cb(); fix this. Fixes: ac888de1510355e341fa8139692b918f2237fa53 https://bugzilla.redhat.com/show_bug.cgi?id=1355740
2016-07-18 18:50:24 +02:00
set_nmc_error_timeout (info->nmc);
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
connection_cb_info_finish (info, NULL);
return G_SOURCE_REMOVE;
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
}
cli: additional fix for nmcli connection down nmcli crashed when a timeout was hit, because 'info' was used after freeing in down_timeout_cb(). Fixes 4a7c88621ddd8e8089f63071e3114003a0efc622.
2014-12-03 13:19:37 +01:00
static void
destroy_queue_element (gpointer data)
{
g_signal_handlers_disconnect_matched (data, G_SIGNAL_MATCH_FUNC, 0, 0, 0,
down_active_connection_state_cb, NULL);
g_object_unref (data);
}
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
static void
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
connection_cb_info_finish (ConnectionCbInfo *info, gpointer connection)
{
if (connection) {
info->queue = g_slist_remove (info->queue, connection);
g_object_unref (G_OBJECT (connection));
cli: additional fix for nmcli connection down nmcli crashed when a timeout was hit, because 'info' was used after freeing in down_timeout_cb(). Fixes 4a7c88621ddd8e8089f63071e3114003a0efc622.
2014-12-03 13:19:37 +01:00
} else {
g_slist_free_full (info->queue, destroy_queue_element);
info->queue = NULL;
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
}
if (info->queue)
return;
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
if (info->timeout_id)
g_source_remove (info->timeout_id);
cli: add missing NULL-check Only the connection down operation is cancellable, the other actions are not. Fixes: 73b560c215806ab72b1473751ba729a3b5596d69
2017-01-29 12:17:43 +01:00
nm_clear_g_cancellable (&info->cancellable);
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
g_signal_handlers_disconnect_by_func (info->nmc->client, connection_removed_cb, info);
g_slice_free (ConnectionCbInfo, info);
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
quit ();
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
static NMCResultCode
do_connection_down (NmCli *nmc, int argc, char **argv)
{
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
NMActiveConnection *active;
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
ConnectionCbInfo *info = NULL;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
const GPtrArray *active_cons;
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
GSList *queue = NULL, *iter;
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
char **arg_arr = NULL;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
char **arg_ptr;
int arg_num;
cli: allow processing all connections with the same name When there are multiple connection profiles of the same name, we used to take and process only the first one. We change the behaviour to process all the connections now in these commands: nmcli connection show <duplicated name> nmcli connection down <duplicated name> nmcli connection delete <duplicated name>
2014-01-14 10:37:25 +01:00
int idx = 0;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
if (nmc->timeout == -1)
nmc->timeout = 10;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
arg_ptr = argv;
arg_num = argc;
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
if (argc == 0) {
cli/connections: add completion to down subcommand
2016-07-12 00:23:27 +02:00
/* nmc_do_cmd() should not call this with argc=0. */
g_assert (!nmc->complete);
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
if (nmc->ask) {
cli: only offer active connections in TAB completion for 'nmcli con down'
2015-03-13 10:02:37 +01:00
char *line = nmc_readline (PROMPT_ACTIVE_CONNECTIONS);
cli: fix TAB completion for multiple connections in 'nmcli con down/delete'
2015-03-12 14:37:47 +01:00
nmc_string_to_arg_array (line, NULL, TRUE, &arg_arr, &arg_num);
cli: fix memory leaks when asking for arguments
2015-03-12 14:42:42 +01:00
g_free (line);
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
arg_ptr = arg_arr;
}
if (arg_num == 0) {
g_string_printf (nmc->return_text, _("Error: No connection specified."));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
cli/connections: add completion to down subcommand
2016-07-12 00:23:27 +02:00
goto finish;
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
}
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
/* Get active connections */
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
active_cons = nm_client_get_active_connections (nmc->client);
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
while (arg_num > 0) {
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
const char *selector = NULL;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
cli/connections: add completion to down subcommand
2016-07-12 00:23:27 +02:00
if (arg_num == 1)
nmc_complete_strings (*arg_ptr, "id", "uuid", "path", "apath", NULL);
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
if ( strcmp (*arg_ptr, "id") == 0
|| strcmp (*arg_ptr, "uuid") == 0
|| strcmp (*arg_ptr, "path") == 0
|| strcmp (*arg_ptr, "apath") == 0) {
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
selector = *arg_ptr;
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
arg_num--;
arg_ptr++;
if (!arg_num) {
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
g_string_printf (nmc->return_text, _("Error: %s argument is missing."), selector);
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
cli/connections: add completion to down subcommand
2016-07-12 00:23:27 +02:00
goto finish;
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
}
}
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
active = find_active_connection (active_cons, connections, selector, *arg_ptr, &idx,
cli/connections: add active connection completion Start completing by the id if the filter type is not specified
2016-07-12 00:23:16 +02:00
arg_num == 1 && nmc->complete);
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
if (active) {
/* Check if the connection is unique. */
/* Calling down for the same connection repeatedly would result in
* NM responding for the last D-Bus call only and we would stall. */
if (!g_slist_find (queue, active))
queue = g_slist_prepend (queue, g_object_ref (active));
} else {
cli/connections: add completion to down subcommand
2016-07-12 00:23:27 +02:00
if (!nmc->complete)
g_printerr (_("Error: '%s' is not an active connection.\n"), *arg_ptr);
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
g_string_printf (nmc->return_text, _("Error: not all active connections found."));
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
}
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
if (idx == 0)
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc->ask ? NULL : nmc, &arg_num, &arg_ptr, NULL);
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
}
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
if (!queue) {
g_string_printf (nmc->return_text, _("Error: no active connection provided."));
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
cli/connections: add completion to down subcommand
2016-07-12 00:23:27 +02:00
goto finish;
} else if (nmc->complete) {
g_slist_free (queue);
goto finish;
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
}
queue = g_slist_reverse (queue);
if (nmc->timeout > 0) {
cli: turn should_wait into a semaphore Count the exit blockers. We'll want to terminate the device monitor in case all monitored devices vanish.
2015-03-27 12:39:17 +01:00
nmc->should_wait++;
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
info = g_slice_new0 (ConnectionCbInfo);
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
info->nmc = nmc;
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
info->queue = queue;
info->timeout_id = g_timeout_add_seconds (nmc->timeout, connection_op_timeout_cb, info);
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
}
for (iter = queue; iter; iter = g_slist_next (iter)) {
active = iter->data;
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
if (info)
cli: wait for "con down" to deactivate the connection (bgo #740775) (rh #1168383) nmcli currently does not wait for the connection to fully deactivate, which can take some time due to dispatcher scripts or cleanup operations like DCB. Change it to wait until the connection is deactivated, or until a short timeout has expired. The user can adjust the timeout with "--wait" if they want. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-11-26 13:23:07 -06:00
g_signal_connect (active,
"notify::" NM_ACTIVE_CONNECTION_STATE,
G_CALLBACK (down_active_connection_state_cb),
info);
cli: make id|uuid optional for 'nmcli connection down' and also allow identifying connetions with 'path' and 'apath' that accepts either the whole D-Dus path or just an index. nmcli connection down [id|uuid|path|apath] <ID> Example: nmcli connection down con1 nmcli connection down con1 con2 apath 1 id my-wifi
2012-11-17 22:42:59 +01:00
cli: fix deactivation for multiple connections (bgo #740775) (rh #1168383) $ nmcli connection down aa bb cc It has been broken by commit 20566c76de0d18162af020852ae1e2055addcfff. Fixups by dcbw. https://bugzilla.gnome.org/show_bug.cgi?id=740775 https://bugzilla.redhat.com/show_bug.cgi?id=1168383
2014-12-02 11:01:53 +01:00
/* Now deactivate the connection */
nm_client_deactivate_connection (nmc->client, active, NULL, NULL);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli/connections: add completion to down subcommand
2016-07-12 00:23:27 +02:00
finish:
cli: 'connection down' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:32:21 +01:00
g_strfreev (arg_arr);
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
return nmc->return_value;
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
/*----------------------------------------------------------------------------*/
typedef struct NameItem {
const char *name;
const char *alias;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
const struct NameItem *settings;
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
gboolean mandatory;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
} NameItem;
cli: allow editing 'generic' connections
2013-11-14 12:58:45 +01:00
static const NameItem nmc_generic_settings [] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
};
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
static const NameItem nmc_ethernet_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, TRUE },
{ NM_SETTING_802_1X_SETTING_NAME, NULL, NULL, FALSE },
cli: add support for DCB settings
2013-11-06 08:16:40 -06:00
{ NM_SETTING_DCB_SETTING_NAME, NULL, NULL, FALSE },
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_infiniband_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_INFINIBAND_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_wifi_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRELESS_SETTING_NAME, "wifi", NULL, TRUE },
{ NM_SETTING_WIRELESS_SECURITY_SETTING_NAME, "wifi-sec", NULL, FALSE },
{ NM_SETTING_802_1X_SETTING_NAME, NULL, NULL, FALSE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_wimax_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIMAX_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_gsm_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_GSM_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_SERIAL_SETTING_NAME, NULL, NULL, FALSE },
cli: allow 'ipv6' and 'ppp' settings for GSM and CDMA connections (rh #1265993) Otherwise the settings cannot be changed. https://bugzilla.redhat.com/show_bug.cgi?id=1265993
2015-09-25 15:42:00 +02:00
{ NM_SETTING_PPP_SETTING_NAME, NULL, NULL, FALSE },
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_cdma_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_CDMA_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_SERIAL_SETTING_NAME, NULL, NULL, FALSE },
cli: allow 'ipv6' and 'ppp' settings for GSM and CDMA connections (rh #1265993) Otherwise the settings cannot be changed. https://bugzilla.redhat.com/show_bug.cgi?id=1265993
2015-09-25 15:42:00 +02:00
{ NM_SETTING_PPP_SETTING_NAME, NULL, NULL, FALSE },
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_bluetooth_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_BLUETOOTH_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_adsl_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_ADSL_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
cli: support PPPoE connection type in nmcli connection editor
2013-12-11 16:54:17 +01:00
/* PPPoE is a base connection type from historical reasons.
all: port everything to libnm Since the API has not changed at this point, this is mostly just a matter of updating Makefiles, and changing references to the library name in comments. NetworkManager cannot link to libnm due to the duplicated type/symbol names. So it links to libnm-core.la directly, which means that NetworkManager gets a separate copy of that code from libnm.so. Everything else links to libnm.
2014-05-19 13:44:02 -04:00
* See libnm-core/nm-setting.c:_nm_setting_is_base_type()
cli: support PPPoE connection type in nmcli connection editor
2013-12-11 16:54:17 +01:00
*/
static const NameItem nmc_pppoe_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
cli: support PPPoE connection type in nmcli connection editor
2013-12-11 16:54:17 +01:00
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, TRUE },
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_PPPOE_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_PPP_SETTING_NAME, NULL, NULL, FALSE },
cli: support PPPoE connection type in nmcli connection editor
2013-12-11 16:54:17 +01:00
{ NM_SETTING_802_1X_SETTING_NAME, NULL, NULL, FALSE },
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_olpc_mesh_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_OLPC_MESH_SETTING_NAME, "olpc-mesh", NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_vpn_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_VPN_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_vlan_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
{ NM_SETTING_VLAN_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_bond_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_BOND_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
cli: add team driver support Signed-off-by: Jiri Pirko <jiri@resnulli.us>
2013-07-25 15:36:49 +02:00
static const NameItem nmc_team_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_TEAM_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
{ NULL, NULL, NULL, FALSE }
cli: add team driver support Signed-off-by: Jiri Pirko <jiri@resnulli.us>
2013-07-25 15:36:49 +02:00
};
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
static const NameItem nmc_bridge_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_BRIDGE_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
static const NameItem nmc_bond_slave_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
cli: add team driver support Signed-off-by: Jiri Pirko <jiri@resnulli.us>
2013-07-25 15:36:49 +02:00
static const NameItem nmc_team_slave_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_TEAM_PORT_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
cli: add team driver support Signed-off-by: Jiri Pirko <jiri@resnulli.us>
2013-07-25 15:36:49 +02:00
};
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
static const NameItem nmc_bridge_slave_settings [] = {
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NM_SETTING_BRIDGE_PORT_SETTING_NAME, NULL, NULL, TRUE },
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
{ NULL, NULL, NULL, FALSE }
};
static const NameItem nmc_no_slave_settings [] = {
{ NM_SETTING_IP4_CONFIG_SETTING_NAME, NULL, NULL, FALSE },
{ NM_SETTING_IP6_CONFIG_SETTING_NAME, NULL, NULL, FALSE },
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
{ NM_SETTING_PROXY_SETTING_NAME, NULL, NULL, FALSE },
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
{ NULL, NULL, NULL, FALSE }
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
};
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
static const NameItem nmc_tun_settings [] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_TUN_SETTING_NAME, NULL, NULL, TRUE },
tun: add support for MTU and cloned-mac-address properties Make it possible to change ethernet.mtu and ethernet.cloned-mac-address properties of tun/tap devices (cloned-mac-address is meaningful only for taps).
2015-12-16 15:08:58 +01:00
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
{ NULL, NULL, NULL, FALSE }
};
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
static const NameItem nmc_ip_tunnel_settings [] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_IP_TUNNEL_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
};
cli: macsec support
2016-07-01 11:56:33 +02:00
static const NameItem nmc_macsec_settings [] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
{ NM_SETTING_802_1X_SETTING_NAME, NULL, NULL, FALSE },
{ NM_SETTING_MACSEC_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
};
cli: add macvlan support
2015-09-18 11:03:17 +02:00
static const NameItem nmc_macvlan_settings [] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
{ NM_SETTING_MACVLAN_SETTING_NAME, NULL, NULL, TRUE },
{ NULL, NULL, NULL, FALSE }
};
cli: add vxlan support
2015-10-13 18:53:10 +02:00
static const NameItem nmc_vxlan_settings [] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_VXLAN_SETTING_NAME, NULL, NULL, TRUE },
vxlan: add support for MTU and cloned-mac-address properties Make it possible to change ethernet.mtu and ethernet.cloned-mac-address properties of vxlan devices.
2015-12-16 14:43:55 +01:00
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
cli: add vxlan support
2015-10-13 18:53:10 +02:00
{ NULL, NULL, NULL, FALSE }
};
cli: support dummy connections
2017-02-01 15:19:29 +01:00
static const NameItem nmc_dummy_settings [] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_DUMMY_SETTING_NAME, NULL, NULL, TRUE },
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", NULL, FALSE },
{ NULL, NULL, NULL, FALSE }
};
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
static const NameItem nmc_valid_connection_types[] = {
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_GENERIC_SETTING_NAME, NULL, nmc_generic_settings }, /* Needs to be first. */
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{ NM_SETTING_WIRED_SETTING_NAME, "ethernet", nmc_ethernet_settings },
cli: support PPPoE connection type in nmcli connection editor
2013-12-11 16:54:17 +01:00
{ NM_SETTING_PPPOE_SETTING_NAME, NULL, nmc_pppoe_settings },
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{ NM_SETTING_WIRELESS_SETTING_NAME, "wifi", nmc_wifi_settings },
{ NM_SETTING_WIMAX_SETTING_NAME, NULL, nmc_wimax_settings },
{ NM_SETTING_GSM_SETTING_NAME, NULL, nmc_gsm_settings },
{ NM_SETTING_CDMA_SETTING_NAME, NULL, nmc_cdma_settings },
{ NM_SETTING_INFINIBAND_SETTING_NAME, NULL, nmc_infiniband_settings },
{ NM_SETTING_ADSL_SETTING_NAME, NULL, nmc_adsl_settings },
{ NM_SETTING_BLUETOOTH_SETTING_NAME, NULL, nmc_bluetooth_settings },
{ NM_SETTING_VPN_SETTING_NAME, NULL, nmc_vpn_settings },
{ NM_SETTING_OLPC_MESH_SETTING_NAME, "olpc-mesh", nmc_olpc_mesh_settings },
{ NM_SETTING_VLAN_SETTING_NAME, NULL, nmc_vlan_settings },
{ NM_SETTING_BOND_SETTING_NAME, NULL, nmc_bond_settings },
cli: add team driver support Signed-off-by: Jiri Pirko <jiri@resnulli.us>
2013-07-25 15:36:49 +02:00
{ NM_SETTING_TEAM_SETTING_NAME, NULL, nmc_team_settings },
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{ NM_SETTING_BRIDGE_SETTING_NAME, NULL, nmc_bridge_settings },
{ "bond-slave", NULL, nmc_bond_slave_settings },
cli: add team driver support Signed-off-by: Jiri Pirko <jiri@resnulli.us>
2013-07-25 15:36:49 +02:00
{ "team-slave", NULL, nmc_team_slave_settings },
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{ "bridge-slave", NULL, nmc_bridge_slave_settings },
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
{ "no-slave", NULL, nmc_no_slave_settings },
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
{ NM_SETTING_TUN_SETTING_NAME, NULL, nmc_tun_settings },
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
{ NM_SETTING_IP_TUNNEL_SETTING_NAME, NULL, nmc_ip_tunnel_settings },
cli: macsec support
2016-07-01 11:56:33 +02:00
{ NM_SETTING_MACSEC_SETTING_NAME, NULL, nmc_macsec_settings },
cli: add macvlan support
2015-09-18 11:03:17 +02:00
{ NM_SETTING_MACVLAN_SETTING_NAME, NULL, nmc_macvlan_settings },
cli: add vxlan support
2015-10-13 18:53:10 +02:00
{ NM_SETTING_VXLAN_SETTING_NAME, NULL, nmc_vxlan_settings },
cli: support dummy connections
2017-02-01 15:19:29 +01:00
{ NM_SETTING_DUMMY_SETTING_NAME, NULL, nmc_dummy_settings },
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{ NULL, NULL, NULL }
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
};
/*
cli: (trivial) fix a typo
2016-12-15 15:27:53 +01:00
* Return the most appropriate name for the connection of a type 'name' possibly with given 'slave_type'
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
* if exists, else return the 'name'. The returned string must not be freed.
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
*/
static const char *
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
get_name_alias (const char *name, const char *slave_type, const NameItem array[])
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
{
const NameItem *iter = &array[0];
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (slave_type) {
while (iter && iter->name) {
if ( g_str_has_prefix (iter->name, slave_type)
&& g_str_has_suffix (iter->name, "-slave"))
break;
iter++;
}
} else if (name) {
while (iter && iter->name) {
if (!strcmp (name, iter->name))
break;
iter++;
}
} else
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
return NULL;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (iter) {
if (iter->alias)
return iter->alias;
else
return iter->name;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
return name;
}
/*
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
* Construct a string with names and aliases from the arrays formatted as:
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
* "name (alias), name, name (alias), name, name"
*
* Returns: string; the caller is responsible for freeing it.
*/
static char *
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
get_valid_options_string (const NameItem *array, const NameItem *array_slv)
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
{
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const NameItem *iter = array;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
GString *str;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
int i;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
str = g_string_sized_new (150);
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
for (i = 0; i < 2; i++, iter = array_slv) {
while (iter && iter->name) {
if (str->len)
g_string_append (str, ", ");
if (iter->alias)
g_string_append_printf (str, "%s (%s)", iter->name, iter->alias);
else
g_string_append (str, iter->name);
iter++;
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
return g_string_free (str, FALSE);
}
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
static const NameItem *
get_valid_settings_array (const char *con_type)
{
guint i, num;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* No connection type yet? Return settings for a generic connection
* (just the "connection" setting), which always makes sense. */
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
if (!con_type)
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return nmc_valid_connection_types[0].settings;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
num = G_N_ELEMENTS (nmc_valid_connection_types);
for (i = 0; i < num; i++) {
if (nm_streq0 (con_type, nmc_valid_connection_types[i].name))
return nmc_valid_connection_types[i].settings;
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
return NULL;
}
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
/* get_valid_properties_string:
* @array: base properties for the current connection type
* @array_slv: slave properties (or ipv4/ipv6 ones) for the current connection type
* @modifier: to prepend to each element of the returned list
* @prefix: only properties matching the prefix will be returned
* @postfix: required prefix on the property args; if a empty string is passed, is
* assumed that the @prefix is a shortcut, so it should not be completed
* but left as is (and an additional check for shortcut ambiguity is performed)
*
* Returns a list of properties compatible with the current connection type
* for the shell autocompletion functionality.
*
* Returns: list of property.arg elements
*/
nmcli: enable connection autocompletion for 802.1x properties nmcli bash autocompletion leveraged on "nmcli connection edit", "print" to retrieve the specific properties of a connection. Anyway, the interactive editor is smart and just prints the used components, so in a connection where 802.1x is not enabled we had no autocompletion. Solved adding an "hidden" command "nmcli --complete connection modify" as suggested in bgo #724860 in order to retrieve ALL the available properties for use in autocompletion. Here patch from L.Rintel has been merged to make che --complete option global to nmcli (first version was local to "connection modify"). https://bugzilla.gnome.org/show_bug.cgi?id=724860 https://bugzilla.redhat.com/show_bug.cgi?id=1301226
2016-06-13 18:59:31 +02:00
static char *
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
get_valid_properties_string (const NameItem *array,
const NameItem *array_slv,
char modifier,
const char *prefix,
const char *postfix)
nmcli: enable connection autocompletion for 802.1x properties nmcli bash autocompletion leveraged on "nmcli connection edit", "print" to retrieve the specific properties of a connection. Anyway, the interactive editor is smart and just prints the used components, so in a connection where 802.1x is not enabled we had no autocompletion. Solved adding an "hidden" command "nmcli --complete connection modify" as suggested in bgo #724860 in order to retrieve ALL the available properties for use in autocompletion. Here patch from L.Rintel has been merged to make che --complete option global to nmcli (first version was local to "connection modify"). https://bugzilla.gnome.org/show_bug.cgi?id=724860 https://bugzilla.redhat.com/show_bug.cgi?id=1301226
2016-06-13 18:59:31 +02:00
{
const NameItem *iter = array;
cli/trivial: rename metadata related names They will be moved out of nmcli as they are generally useful. Even if they happen to be used only inside nmcli, there should be a better separation between logic (nmcli) and setting decriptions.
2017-03-28 11:02:03 +02:00
const NMMetaSettingInfoEditor *setting_info;
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
const char *prop_name = NULL;
nmcli: enable connection autocompletion for 802.1x properties nmcli bash autocompletion leveraged on "nmcli connection edit", "print" to retrieve the specific properties of a connection. Anyway, the interactive editor is smart and just prints the used components, so in a connection where 802.1x is not enabled we had no autocompletion. Solved adding an "hidden" command "nmcli --complete connection modify" as suggested in bgo #724860 in order to retrieve ALL the available properties for use in autocompletion. Here patch from L.Rintel has been merged to make che --complete option global to nmcli (first version was local to "connection modify"). https://bugzilla.gnome.org/show_bug.cgi?id=724860 https://bugzilla.redhat.com/show_bug.cgi?id=1301226
2016-06-13 18:59:31 +02:00
GString *str;
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
int i, j;
nmcli: fix ambiguous properties check for autocompletion Re-enable the check on ambiguos properties but this time manage perfect matches (i.e., the full property name has been typed) that are prefix of other properties. Test case: nmcli --complete-args connection add type wifi wifi. Here "wifi" is a property alias that is also prefix of the property alias "wifi-sec".
2016-06-23 17:17:10 +02:00
gboolean full_match = FALSE;
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
g_return_val_if_fail (prefix, NULL);
nmcli: enable connection autocompletion for 802.1x properties nmcli bash autocompletion leveraged on "nmcli connection edit", "print" to retrieve the specific properties of a connection. Anyway, the interactive editor is smart and just prints the used components, so in a connection where 802.1x is not enabled we had no autocompletion. Solved adding an "hidden" command "nmcli --complete connection modify" as suggested in bgo #724860 in order to retrieve ALL the available properties for use in autocompletion. Here patch from L.Rintel has been merged to make che --complete option global to nmcli (first version was local to "connection modify"). https://bugzilla.gnome.org/show_bug.cgi?id=724860 https://bugzilla.redhat.com/show_bug.cgi?id=1301226
2016-06-13 18:59:31 +02:00
str = g_string_sized_new (1024);
for (i = 0; i < 2; i++, iter = array_slv) {
nmcli: fix ambiguous properties check for autocompletion Re-enable the check on ambiguos properties but this time manage perfect matches (i.e., the full property name has been typed) that are prefix of other properties. Test case: nmcli --complete-args connection add type wifi wifi. Here "wifi" is a property alias that is also prefix of the property alias "wifi-sec".
2016-06-23 17:17:10 +02:00
while (!full_match && iter && iter->name) {
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
if ( !(g_str_has_prefix (iter->name, prefix))
&& (!(iter->alias) || !g_str_has_prefix (iter->alias, prefix))) {
iter++;
continue;
}
/* If postix (so prefix is terminated by a dot), check
* that prefix is not ambiguous */
if (postfix) {
nmcli: fix ambiguous properties check for autocompletion Re-enable the check on ambiguos properties but this time manage perfect matches (i.e., the full property name has been typed) that are prefix of other properties. Test case: nmcli --complete-args connection add type wifi wifi. Here "wifi" is a property alias that is also prefix of the property alias "wifi-sec".
2016-06-23 17:17:10 +02:00
/* If we have a perfect match, no need to look for others
* prefix and no check on ambiguity should be performed.
* Moreover, erase previous matches from output string */
if ( nm_streq (prefix, iter->name)
|| nm_streq0 (prefix, iter->alias)) {
g_string_erase (str, 0, -1);
full_match = TRUE;
} else if (prop_name) {
return g_string_free (str, TRUE);
}
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
prop_name = prefix;
} else {
prop_name = iter->name;
}
nmcli: enable connection autocompletion for 802.1x properties nmcli bash autocompletion leveraged on "nmcli connection edit", "print" to retrieve the specific properties of a connection. Anyway, the interactive editor is smart and just prints the used components, so in a connection where 802.1x is not enabled we had no autocompletion. Solved adding an "hidden" command "nmcli --complete connection modify" as suggested in bgo #724860 in order to retrieve ALL the available properties for use in autocompletion. Here patch from L.Rintel has been merged to make che --complete option global to nmcli (first version was local to "connection modify"). https://bugzilla.gnome.org/show_bug.cgi?id=724860 https://bugzilla.redhat.com/show_bug.cgi?id=1301226
2016-06-13 18:59:31 +02:00
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
/* Search the array with the arguments of the current property */
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
setting_info = nm_meta_setting_info_editor_find_by_name (iter->name);
nmcli: enable connection autocompletion for 802.1x properties nmcli bash autocompletion leveraged on "nmcli connection edit", "print" to retrieve the specific properties of a connection. Anyway, the interactive editor is smart and just prints the used components, so in a connection where 802.1x is not enabled we had no autocompletion. Solved adding an "hidden" command "nmcli --complete connection modify" as suggested in bgo #724860 in order to retrieve ALL the available properties for use in autocompletion. Here patch from L.Rintel has been merged to make che --complete option global to nmcli (first version was local to "connection modify"). https://bugzilla.gnome.org/show_bug.cgi?id=724860 https://bugzilla.redhat.com/show_bug.cgi?id=1301226
2016-06-13 18:59:31 +02:00
j = 0;
cli: rework property handling via NmcPropertyInfo for NMSettingConnection Add an improved way of tracking meta data about settings. E.g. it is wrong to generate for each property a nmc_property_*_get_*() function. Instead, there should be a meta data about the property itself, and a mechanism to retrieve the property. For now, only do this for NMSettingConnection and keep all the existing infrastructure in place. Later on all settings shall be moved. Especially to accomodate NmcOutputField mangles the concept of setting-meta data, formatting options, and collecting output results. It's a total hack, that will be need fixing later.
2017-03-24 17:32:04 +01:00
while (TRUE) {
nmcli: improve connection autocompletion (1/2) Make property autocompletion take a prefix and modifier flags. This will make it easier to complete an unfinished property name (possibly accompanied by a modifier) without shell trickery.
2016-06-15 13:10:26 +02:00
gchar *new;
cli: rework property handling via NmcPropertyInfo for NMSettingConnection Add an improved way of tracking meta data about settings. E.g. it is wrong to generate for each property a nmc_property_*_get_*() function. Instead, there should be a meta data about the property itself, and a mechanism to retrieve the property. For now, only do this for NMSettingConnection and keep all the existing infrastructure in place. Later on all settings shall be moved. Especially to accomodate NmcOutputField mangles the concept of setting-meta data, formatting options, and collecting output results. It's a total hack, that will be need fixing later.
2017-03-24 17:32:04 +01:00
const char *arg_name;
cli: split tracking of meta data out of NmcOutputField When generating output data, nmcli iterates over a list of property-descriptors (nmc_fields_ip4_config), creates an intermediate array (output_data) and finally prints it. However, previously both the meta data (nmc_fields_ip4_config) and the intermediate format use the same type NmcOutputField. This means, certain fields are relevant to describe a property, and other fields are output/formatting fields. Split this up. Now, the meta data is tracked in form of an NMMetaAbstractInfo lists. This separates the information about properties from intermediate steps during creation of the output. Note that currently functions like print_ip4_config() still have all the knowledge about how to generate the output. That is wrong, instead, the meta data (NMMetaAbstractInfo) should describe how to create the output and then all those functions could be replaced. This means, later we want to add more knowledge to the NMMetaAbstractInfo, so it is important to keep them separate from NmcOutputField.
2017-03-31 19:18:16 +02:00
if (j + 1 >= setting_info->properties_num)
break;
arg_name = setting_info->properties[j + 1].property_name;
nmcli: improve connection autocompletion (1/2) Make property autocompletion take a prefix and modifier flags. This will make it easier to complete an unfinished property name (possibly accompanied by a modifier) without shell trickery.
2016-06-15 13:10:26 +02:00
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
/* If required, expand the alias too */
if (!postfix && iter->alias) {
nmcli: improve connection autocompletion (1/2) Make property autocompletion take a prefix and modifier flags. This will make it easier to complete an unfinished property name (possibly accompanied by a modifier) without shell trickery.
2016-06-15 13:10:26 +02:00
if (modifier)
g_string_append_c (str, modifier);
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
new = g_strdup_printf ("%s.%s\n",
cli: rework property handling via NmcPropertyInfo for NMSettingConnection Add an improved way of tracking meta data about settings. E.g. it is wrong to generate for each property a nmc_property_*_get_*() function. Instead, there should be a meta data about the property itself, and a mechanism to retrieve the property. For now, only do this for NMSettingConnection and keep all the existing infrastructure in place. Later on all settings shall be moved. Especially to accomodate NmcOutputField mangles the concept of setting-meta data, formatting options, and collecting output results. It's a total hack, that will be need fixing later.
2017-03-24 17:32:04 +01:00
iter->alias,
arg_name);
nmcli: improve connection autocompletion (1/2) Make property autocompletion take a prefix and modifier flags. This will make it easier to complete an unfinished property name (possibly accompanied by a modifier) without shell trickery.
2016-06-15 13:10:26 +02:00
g_string_append (str, new);
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
g_free (new);
nmcli: improve connection autocompletion (1/2) Make property autocompletion take a prefix and modifier flags. This will make it easier to complete an unfinished property name (possibly accompanied by a modifier) without shell trickery.
2016-06-15 13:10:26 +02:00
}
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
if (postfix && !g_str_has_prefix (arg_name, postfix)) {
j++;
continue;
nmcli: improve connection autocompletion (1/2) Make property autocompletion take a prefix and modifier flags. This will make it easier to complete an unfinished property name (possibly accompanied by a modifier) without shell trickery.
2016-06-15 13:10:26 +02:00
}
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
if (modifier)
g_string_append_c (str, modifier);
new = g_strdup_printf ("%s.%s\n",
cli: rework property handling via NmcPropertyInfo for NMSettingConnection Add an improved way of tracking meta data about settings. E.g. it is wrong to generate for each property a nmc_property_*_get_*() function. Instead, there should be a meta data about the property itself, and a mechanism to retrieve the property. For now, only do this for NMSettingConnection and keep all the existing infrastructure in place. Later on all settings shall be moved. Especially to accomodate NmcOutputField mangles the concept of setting-meta data, formatting options, and collecting output results. It's a total hack, that will be need fixing later.
2017-03-24 17:32:04 +01:00
prop_name,
arg_name);
nmcli: add support to shortnames for connection properties autocompletion
2016-06-18 21:17:57 +02:00
g_string_append (str, new);
g_free (new);
nmcli: enable connection autocompletion for 802.1x properties nmcli bash autocompletion leveraged on "nmcli connection edit", "print" to retrieve the specific properties of a connection. Anyway, the interactive editor is smart and just prints the used components, so in a connection where 802.1x is not enabled we had no autocompletion. Solved adding an "hidden" command "nmcli --complete connection modify" as suggested in bgo #724860 in order to retrieve ALL the available properties for use in autocompletion. Here patch from L.Rintel has been merged to make che --complete option global to nmcli (first version was local to "connection modify"). https://bugzilla.gnome.org/show_bug.cgi?id=724860 https://bugzilla.redhat.com/show_bug.cgi?id=1301226
2016-06-13 18:59:31 +02:00
j++;
}
iter++;
}
}
return g_string_free (str, FALSE);
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
/*
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
* Check if 'val' is valid string in either array->name or array->alias for
* both array parameters (array & array_slv).
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
* It accepts shorter string provided they are not ambiguous.
* 'val' == NULL doesn't hurt.
*
* Returns: pointer to array->name string or NULL on failure.
* The returned string must not be freed.
*/
static const char *
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
check_valid_name (const char *val, const NameItem *array, const NameItem *array_slv, GError **error)
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
{
const NameItem *iter;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
gs_unref_ptrarray GPtrArray *tmp_arr = NULL;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
const char *str;
GError *tmp_err = NULL;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
int i;
g_return_val_if_fail (array, NULL);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
/* Create a temporary array that can be used in nmc_string_is_valid() */
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
tmp_arr = g_ptr_array_sized_new (32);
iter = array;
for (i = 0; i < 2; i++, iter = array_slv) {
while (iter && iter->name) {
g_ptr_array_add (tmp_arr, (gpointer) iter->name);
if (iter->alias)
g_ptr_array_add (tmp_arr, (gpointer) iter->alias);
iter++;
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
g_ptr_array_add (tmp_arr, (gpointer) NULL);
/* Check string validity */
str = nmc_string_is_valid (val, (const char **) tmp_arr->pdata, &tmp_err);
if (!str) {
if (tmp_err->code == 1)
g_propagate_error (error, tmp_err);
else {
/* We want to handle aliases, so construct own error message */
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
char *err_str = get_valid_options_string (array, array_slv);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
g_set_error (error, 1, 0, _("'%s' not among [%s]"), val, err_str);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
g_free (err_str);
g_clear_error (&tmp_err);
}
return NULL;
}
/* Return a pointer to the found string in passed 'array' */
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
iter = array;
for (i = 0; i < 2; i++, iter = array_slv) {
while (iter && iter->name) {
if ( nm_streq (iter->name, str)
|| nm_streq0 (iter->alias, str)) {
return iter->name;
}
iter++;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
}
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
/* We should not really come here */
g_set_error (error, 1, 0, _("Unknown error"));
return NULL;
}
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
static gboolean
is_setting_mandatory (NMConnection *connection, NMSetting *setting)
{
NMSettingConnection *s_con;
const char *c_type;
const NameItem *item;
const char *name;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const char *s_type;
char *slv_type;
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
s_con = nm_connection_get_setting_connection (connection);
g_assert (s_con);
c_type = nm_setting_connection_get_connection_type (s_con);
name = nm_setting_get_name (setting);
item = get_valid_settings_array (c_type);
while (item && item->name) {
if (!strcmp (name, item->name))
return item->mandatory;
item++;
}
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
/* Let's give a try to parameters related to slave type */
s_type = nm_setting_connection_get_slave_type (s_con);
slv_type = g_strdup_printf ("%s-slave", s_type ? s_type : "no");
item = get_valid_settings_array (slv_type);
g_free (slv_type);
while (item && item->name) {
if (!strcmp (name, item->name))
return item->mandatory;
item++;
}
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
return FALSE;
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
/*----------------------------------------------------------------------------*/
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
static const char *
_strip_master_prefix (const char *master, const char *(**func)(NMConnection *))
{
if (!master)
return NULL;
if (g_str_has_prefix (master, "ifname/")) {
master = master + strlen ("ifname/");
if (func)
libnm-core: drop nm_{setting,connection}_get_virtual_iface_name() Since we enforce the fact that bond, bridge, team, and vlan interface-name properties match NMSettingConnection:interface-name, nm_connection_get_virtual_iface_name() can be replaced with nm_connection_get_interface_name() basically everywhere. The one place this doesn't work is with InfiniBand partitions (where get_virtual_iface_name() was actually computing the name), but for the most part we only need to care about the interface names of InfiniBand partitions in places where we also already need to do some other InfiniBand-specific handling as well, so we can use an InfiniBand-specific method (nm_setting_infiniband_get_virtual_interface_name()) to get it. (Also, while updating nm_device_get_virtual_device_description(), fix it to handle InfiniBand partitions too.)
2014-08-07 10:47:56 -04:00
*func = nm_connection_get_interface_name;
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
} else if (g_str_has_prefix (master, "uuid/")) {
master = master + strlen ("uuid/");
if (func)
*func = nm_connection_get_uuid;
} else if (g_str_has_prefix (master, "id/")) {
master = master + strlen ("id/");
if (func)
*func = nm_connection_get_id;
}
return master;
}
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
/* normalized_master_for_slave:
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
* @connections: list af all connections
* @master: UUID, ifname or ID of the master connection
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
* @type: virtual connection type (bond, team, bridge, ...) or %NULL
* @out_type: type of the connection that matched
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
*
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
* Check whether master is a valid interface name, UUID or ID of some connection,
* possibly of a specified @type.
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
* First UUID and ifname are checked. If they don't match, ID is checked
* and replaced by UUID on a match.
*
* Returns: identifier of master connection if found, %NULL otherwise
*/
static const char *
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
normalized_master_for_slave (const GPtrArray *connections,
const char *master,
const char *type,
const char **out_type)
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
{
NMConnection *connection;
NMSettingConnection *s_con;
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
const char *con_type = NULL, *id, *uuid, *ifname;
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
int i;
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
const char *found_by_id = NULL;
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
const char *out_type_by_id = NULL;
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
const char *out_master = NULL;
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
const char *(*func) (NMConnection *) = NULL;
if (!master)
return NULL;
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
master = _strip_master_prefix (master, &func);
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
for (i = 0; i < connections->len; i++) {
connection = NM_CONNECTION (connections->pdata[i]);
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
s_con = nm_connection_get_setting_connection (connection);
g_assert (s_con);
con_type = nm_setting_connection_get_connection_type (s_con);
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
if (type && g_strcmp0 (con_type, type) != 0)
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
continue;
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
if (func) {
/* There was a prefix; only compare to that type. */
if (g_strcmp0 (master, func (connection)) == 0) {
cli: fix a crash adding a slave connection with 'nmcli -a con add' Valid types: [generic, 802-3-ethernet (ethernet), pppoe, 802-11-wireless (wifi), wimax, gsm, cdma, infiniband, adsl, bluetooth, vpn, 802-11-olpc-mesh (olpc-mesh), vlan, bond, team, bridge, bond-slave, team-slave, bridge-slave] Connection type: team-slave Interface name [*]: eth5 Master: nm-team Program received signal SIGSEGV, Segmentation fault. 0x000000000041ae6d in normalized_master_for_slave (connections=connections@entry=0x7fffec0019a0, master=<optimized out>, master@entry=0x73b470 "nm-team", type=type@entry=0x72f7c0 "team", out_type=out_type@entry=0x0) at connections.c:3218 3218 *out_type = con_type; Fixes: aa12bb353bca34be1bea0625c8e6e7715f24deb3
2015-07-15 17:16:11 +02:00
if (out_type)
*out_type = con_type;
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
if (func == nm_connection_get_id)
out_master = nm_connection_get_uuid (connection);
else
out_master = master;
break;
}
} else {
id = nm_connection_get_id (connection);
uuid = nm_connection_get_uuid (connection);
libnm-core: drop nm_{setting,connection}_get_virtual_iface_name() Since we enforce the fact that bond, bridge, team, and vlan interface-name properties match NMSettingConnection:interface-name, nm_connection_get_virtual_iface_name() can be replaced with nm_connection_get_interface_name() basically everywhere. The one place this doesn't work is with InfiniBand partitions (where get_virtual_iface_name() was actually computing the name), but for the most part we only need to care about the interface names of InfiniBand partitions in places where we also already need to do some other InfiniBand-specific handling as well, so we can use an InfiniBand-specific method (nm_setting_infiniband_get_virtual_interface_name()) to get it. (Also, while updating nm_device_get_virtual_device_description(), fix it to handle InfiniBand partitions too.)
2014-08-07 10:47:56 -04:00
ifname = nm_connection_get_interface_name (connection);
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
if ( g_strcmp0 (master, uuid) == 0
|| g_strcmp0 (master, ifname) == 0) {
out_master = master;
cli: fix a crash adding a slave connection with 'nmcli -a con add' Valid types: [generic, 802-3-ethernet (ethernet), pppoe, 802-11-wireless (wifi), wimax, gsm, cdma, infiniband, adsl, bluetooth, vpn, 802-11-olpc-mesh (olpc-mesh), vlan, bond, team, bridge, bond-slave, team-slave, bridge-slave] Connection type: team-slave Interface name [*]: eth5 Master: nm-team Program received signal SIGSEGV, Segmentation fault. 0x000000000041ae6d in normalized_master_for_slave (connections=connections@entry=0x7fffec0019a0, master=<optimized out>, master@entry=0x73b470 "nm-team", type=type@entry=0x72f7c0 "team", out_type=out_type@entry=0x0) at connections.c:3218 3218 *out_type = con_type; Fixes: aa12bb353bca34be1bea0625c8e6e7715f24deb3
2015-07-15 17:16:11 +02:00
if (out_type)
*out_type = con_type;
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
break;
}
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
if (!found_by_id && g_strcmp0 (master, id) == 0) {
out_type_by_id = con_type;
cli: accept prefix "ifname/", "uuid/" or "id/" for 'master' argument nmcli con add type team-slave ifname em2 master team-master0 nmcli con add type team-slave ifname em2 master id/team-master0 It helps to disambiguate values for cases where they may overlap, e.g. "team0" -> "ifname/team0" or "id/team0"
2014-03-11 13:21:50 +01:00
found_by_id = uuid;
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
}
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
}
}
cli: discover slave type for a connection with a master Rename verify_master_for_slave(), since it does a lot more than just verifying the master setting. Make the type check optional and return the type of the connection that matched. This makes it possible to omit setting the slave type on a command line and still get the slave type right.
2015-04-21 17:26:35 +02:00
if (!out_master) {
out_master = found_by_id;
if (out_type)
*out_type = out_type_by_id;
}
if (!out_master) {
g_print (_("Warning: master='%s' doesn't refer to any existing profile.\n"), master);
out_master = master;
if (out_type)
*out_type = type;
}
return out_master;
cli: allow specifying 'master' for slaves as connection ID (rh #1057494) nmcli con add type *-slave ifname em1 master <ifname|UUID|name> 'master' property of 'connection' setting has to be either interface name or connection UUID of master connection. However, to make nmcli more convenient for users, we also allow specifying connection name and translating it to UUID automatically. https://bugzilla.redhat.com/show_bug.cgi?id=1057494
2014-03-11 12:10:00 +01:00
}
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
#define WORD_YES "yes"
#define WORD_NO "no"
#define WORD_LOC_YES _("yes")
#define WORD_LOC_NO _("no")
static const char *
prompt_yes_no (gboolean default_yes, char *delim)
{
static char prompt[128] = { 0 };
if (!delim)
delim = "";
snprintf (prompt, sizeof (prompt), "(%s/%s) [%s]%s ",
WORD_LOC_YES, WORD_LOC_NO,
default_yes ? WORD_LOC_YES : WORD_LOC_NO, delim);
return prompt;
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static NMSetting *
is_setting_valid (NMConnection *connection, const NameItem *valid_settings_main, const NameItem *valid_settings_slave, char *setting)
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *setting_name;
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!(setting_name = check_valid_name (setting, valid_settings_main, valid_settings_slave, NULL)))
return NULL;
return nm_connection_get_setting_by_name (connection, setting_name);
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
is_property_valid (NMSetting *setting, const char *property, GError **error)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
char **valid_props = NULL;
const char *prop_name;
char *ret;
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
valid_props = nmc_setting_get_valid_properties (setting);
prop_name = nmc_string_is_valid (property, (const char **) valid_props, error);
ret = g_strdup (prop_name);
g_strfreev (valid_props);
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
return ret;
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
unique_master_iface_ifname (const GPtrArray *connections,
const char *try_name)
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
NMConnection *connection;
char *new_name;
all: use "unsigned" instead of "unsigned int"
2017-03-14 11:15:05 +01:00
unsigned num = 1;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
int i = 0;
const char *ifname = NULL;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
new_name = g_strdup (try_name);
while (i < connections->len) {
connection = NM_CONNECTION (connections->pdata[i]);
ifname = nm_connection_get_interface_name (connection);
if (g_strcmp0 (new_name, ifname) == 0) {
g_free (new_name);
new_name = g_strdup_printf ("%s%d", try_name, num++);
i = 0;
} else
i++;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return new_name;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
static void
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
set_default_interface_name (NmCli *nmc, NMSettingConnection *s_con)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
char *ifname = NULL;
const char *con_type = nm_setting_connection_get_connection_type (s_con);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (nm_setting_connection_get_interface_name (s_con))
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
return;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* Set a sensible bond/team/bridge interface name by default */
if (g_strcmp0 (con_type, NM_SETTING_BOND_SETTING_NAME) == 0)
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
ifname = unique_master_iface_ifname (connections, "nm-bond");
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_strcmp0 (con_type, NM_SETTING_TEAM_SETTING_NAME) == 0)
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
ifname = unique_master_iface_ifname (connections, "nm-team");
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_strcmp0 (con_type, NM_SETTING_BRIDGE_SETTING_NAME) == 0)
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
ifname = unique_master_iface_ifname (connections, "nm-bridge");
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else
cli: Add mode option for wifi connecitons Only default (infrastructure) mode connections can be created and as it's not possible to write mode=ap connections with ifcfg-rh plugins, they can't be switched to mode=ap.
2014-10-04 17:07:42 +02:00
return;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
g_object_set (s_con, NM_SETTING_CONNECTION_INTERFACE_NAME, ifname, NULL);
g_free (ifname);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*----------------------------------------------------------------------------*/
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static OptionInfo option_info[];
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*
* Mark options in option_info as relevant.
* The questionnaire (for --ask) will ask for them.
*/
cli: support PPPoE connection types in 'nmcli con add'
2013-12-11 17:55:19 +01:00
static void
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
enable_options (const gchar *setting_name, const gchar *property, const gchar * const *opts)
cli: support PPPoE connection types in 'nmcli con add'
2013-12-11 17:55:19 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
OptionInfo *candidate;
cli: support PPPoE connection types in 'nmcli con add'
2013-12-11 17:55:19 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
for (candidate = option_info; candidate->setting_name; candidate++) {
if ( strcmp (candidate->setting_name, setting_name) == 0
&& strcmp (candidate->property, property) == 0
&& (candidate->flags & OPTION_DONT_ASK)
&& candidate->option
&& g_strv_contains (opts, candidate->option)) {
candidate->flags |= OPTION_ENABLED;
}
cli: support PPPoE connection types in 'nmcli con add'
2013-12-11 17:55:19 +01:00
}
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*
* Mark options in option_info as irrelevant (because we learned they make no sense
* or they have been set via different means).
* The questionnaire (for --ask) will not ask for them.
*/
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
static void
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
disable_options (const gchar *setting_name, const gchar *property)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
OptionInfo *candidate;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
for (candidate = option_info; candidate->setting_name; candidate++) {
if ( strcmp (candidate->setting_name, setting_name) == 0
&& (!property || strcmp (candidate->property, property) == 0))
candidate->flags |= OPTION_DISABLED;
}
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*
* Reset marks done with enable_options() and disable_options().
* Ensures correct operation in case more than one connection is added in a single
* nmcli session.
*/
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
static void
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
reset_options (void)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
OptionInfo *candidate;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
for (candidate = option_info; candidate->setting_name; candidate++) {
candidate->flags &= ~OPTION_DISABLED;
candidate->flags &= ~OPTION_ENABLED;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_property (NMConnection *connection,
const char *setting_name, const char *property, const char *value,
char modifier, GError **error)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
gs_free char *property_name = NULL, *value_free = NULL;
NMSetting *setting;
GError *local = NULL;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
setting = nm_connection_get_setting_by_name (connection, setting_name);
if (!setting) {
setting = nmc_setting_new_for_name (setting_name);
if (!setting) {
/* This should really not happen */
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_UNKNOWN,
_("Error: don't know how to create '%s' setting."),
setting_name);
return FALSE;
}
nmc_setting_custom_init (setting);
nm_connection_add_setting (connection, setting);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
property_name = is_property_valid (setting, property, &local);
if (!property_name) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: invalid property '%s': %s."),
property, local->message);
g_clear_error (&local);
return FALSE;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (modifier != '-') {
/* Set/add value */
if (modifier != '+') {
/* We allow the existing property value to be passed as parameter,
* so make a copy if we are going to free it.
*/
value = value_free = g_strdup (value);
nmc_setting_reset_property (setting, property_name, NULL);
}
if (!nmc_setting_set_property (setting, property_name, value, &local)) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: failed to modify %s.%s: %s."),
setting_name, property, local->message);
g_clear_error (&local);
return FALSE;
}
} else {
/* Remove value
* - either empty: remove whole value
* - or specified by index <0-n>: remove item at the index
* - or option name: remove item with the option name
*/
if (value) {
unsigned long idx;
if (nmc_string_to_uint (value, TRUE, 0, G_MAXUINT32, &idx))
nmc_setting_remove_property_option (setting, property_name, NULL, idx, &local);
else
nmc_setting_remove_property_option (setting, property_name, value, 0, &local);
if (local) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: failed to remove a value from %s.%s: %s."),
setting_name, property, local->message);
g_clear_error (&local);
return FALSE;
cli: fix memory leaks on questionnaires
2013-10-01 18:45:26 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
} else
nmc_setting_reset_property (setting, property_name, NULL);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* Don't ask for this property in interactive mode. */
disable_options (setting_name, property_name);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return TRUE;
}
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_option (NmCli *nmc, NMConnection *connection, OptionInfo *option, const gchar *value, GError **error)
{
option->flags |= OPTION_DISABLED;
if (option->check_and_set) {
return option->check_and_set (nmc, connection, option, value, error);
} else if (value) {
return set_property (connection, option->setting_name, option->property,
value, option->flags & OPTION_MULTI ? '+' : '\0', error);
} else if (option->flags & OPTION_REQD) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: '%s' is mandatory."), option->option);
return FALSE;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
cli: enhance questionnaire for bond connections (rh #1007355) The session now looks like this: There are optional arguments for 'bond' connection type. Do you want to provide them? (yes/no) [yes] Bonding mode [balance-rr]: Bonding monitoring mode (miimon or arp) [miimon]: miimon Bonding miimon [100]: Bonding downdelay [0]: Bonding updelay [0]: ...when answered arp... Bonding arp-interval [0]: Bonding arp-ip-target [none]: https://bugzilla.redhat.com/show_bug.cgi?id=1007355
2013-10-01 18:46:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return TRUE;
}
cli: accept 'primary' bonding option
2013-10-01 22:32:48 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*
* Return relevant NameItem[] tables for given connection (based on connection type
* and slave type.
*/
static gboolean
con_settings (NMConnection *connection, const NameItem **type_settings, const NameItem **slv_settings, GError **error)
{
const char *con_type;
gs_free char *slv_type = NULL;
NMSettingConnection *s_con;
g_return_val_if_fail (type_settings, FALSE);
g_return_val_if_fail (slv_settings, FALSE);
s_con = nm_connection_get_setting_connection (connection);
g_assert (s_con);
con_type = nm_setting_connection_get_slave_type (s_con);
if (!con_type)
con_type = "no";
slv_type = g_strdup_printf ("%s-slave", con_type);
if (slv_type) {
*slv_settings = get_valid_settings_array (slv_type);
if (!*slv_settings) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: invalid slave type; %s."), slv_type);
return FALSE;
cli: enhance questionnaire for bond connections (rh #1007355) The session now looks like this: There are optional arguments for 'bond' connection type. Do you want to provide them? (yes/no) [yes] Bonding mode [balance-rr]: Bonding monitoring mode (miimon or arp) [miimon]: miimon Bonding miimon [100]: Bonding downdelay [0]: Bonding updelay [0]: ...when answered arp... Bonding arp-interval [0]: Bonding arp-ip-target [none]: https://bugzilla.redhat.com/show_bug.cgi?id=1007355
2013-10-01 18:46:10 +02:00
}
} else {
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
*slv_settings = NULL;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
con_type = nm_setting_connection_get_connection_type (s_con);
if (!con_type)
con_type = NM_SETTING_GENERIC_SETTING_NAME;
*type_settings = get_valid_settings_array (con_type);
if (!*type_settings) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: invalid connection type; %s."), con_type);
return FALSE;
cli: support new bond option 'lacp_rate' Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-25 12:59:49 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return TRUE;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*
* Make sure all required settings are in place (should be called when
* it's possible that a type is aready set).
*/
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
static void
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
ensure_settings (NMConnection *connection, const NameItem *item)
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const NameItem *setting_item;
NMSetting *setting;
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
for (setting_item = item; setting_item->name; setting_item++) {
if (!setting_item->mandatory)
continue;
if (nm_connection_get_setting_by_name (connection, setting_item->name))
continue;
setting = nmc_setting_new_for_name (setting_item->name);
if (setting) {
nmc_setting_custom_init (setting);
nm_connection_add_setting (connection, setting);
}
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
}
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*----------------------------------------------------------------------------*/
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
static char *
gen_func_slave_type (const char *text, int state)
{
const char *words[] = { "bond", "team", "bridge", NULL };
return nmc_rl_gen_func_basic (text, state, words);
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_vpn_types (const char *text, int state)
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
gs_strfreev char **plugin_names = NULL;
plugin_names = nm_vpn_plugin_info_list_get_service_types (nm_vpn_get_plugin_infos (), FALSE, TRUE);
return nmc_rl_gen_func_basic (text, state, (const char **) plugin_names);
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_bool_values_l10n (const char *text, int state)
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { WORD_LOC_YES, WORD_LOC_NO, NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: support file names for 'config' argument when creating team connections nmcli con add type team config /home/cimrman/team-config.json libteam (and in turn NetworkManager) configures team devices via plain config data in JSON format. However, it is useful and more user-friendly for nmcli to accept also a file name that contains the config data, and read it. Thus the user is not forced to type whole (possibly long) config on the command line.
2013-11-11 17:48:25 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_wifi_mode (const char *text, int state)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "infrastructure", "ap", "adhoc", NULL };
return nmc_rl_gen_func_basic (text, state, words);
}
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_ib_type (const char *text, int state)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "datagram", "connected", NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_bt_type (const char *text, int state)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "panu", "dun-gsm", "dun-cdma", NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_bond_mode (const char *text, int state)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "balance-rr", "active-backup", "balance-xor", "broadcast",
"802.3ad", "balance-tlb", "balance-alb", NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_bond_mon_mode (const char *text, int state)
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "miimon", "arp", NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
}
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
static char *
gen_func_bond_lacp_rate (const char *text, int state)
{
const char *words[] = { "slow", "fast", NULL };
return nmc_rl_gen_func_basic (text, state, words);
}
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_adsl_proto (const char *text, int state)
cli: add macvlan support
2015-09-18 11:03:17 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "pppoe", "pppoa", "ipoatm", NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: add macvlan support
2015-09-18 11:03:17 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_adsl_encap (const char *text, int state)
cli: add vxlan support
2015-10-13 18:53:10 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "vcmux", "llc", NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: add vxlan support
2015-10-13 18:53:10 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_tun_mode (const char *text, int state)
cli: accept gateway in the IP questionnaire of 'nmcli -a con add' (rh #1007368) Also allow leading/trailing white spaces, and more white spaces between IP and gateway. (Spaces in values and around / are not allowed). All of these are accepted: ' 1.2.3.4/11 1.2.3.254 ' ' 1.2.3.4 ' ' fe80::215:ff:fe93:ffff/128 ::1 ' ... https://bugzilla.redhat.com/show_bug.cgi?id=1007368
2013-09-19 11:39:40 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const char *words[] = { "tun", "tap", NULL };
return nmc_rl_gen_func_basic (text, state, words);
cli: accept gateway in the IP questionnaire of 'nmcli -a con add' (rh #1007368) Also allow leading/trailing white spaces, and more white spaces between IP and gateway. (Spaces in values and around / are not allowed). All of these are accepted: ' 1.2.3.4/11 1.2.3.254 ' ' 1.2.3.4 ' ' fe80::215:ff:fe93:ffff/128 ::1 ' ... https://bugzilla.redhat.com/show_bug.cgi?id=1007368
2013-09-19 11:39:40 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_ip_tunnel_mode (const char *text, int state)
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
gs_free const char **words = NULL;
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
words = nm_utils_enum_get_values (nm_ip_tunnel_mode_get_type (),
NM_IP_TUNNEL_MODE_UNKNOWN + 1,
G_MAXINT);
return nmc_rl_gen_func_basic (text, state, words);
cli: use a generic function asking for both IPv4/IPv6 in questionnaire
2013-09-19 13:54:04 +02:00
}
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
cli: macsec support
2016-07-01 11:56:33 +02:00
static char *
gen_func_macsec_mode (const char *text, int state)
{
gs_free const char **words = NULL;
words = nm_utils_enum_get_values (nm_setting_macsec_mode_get_type (),
G_MININT,
G_MAXINT);
return nmc_rl_gen_func_basic (text, state, words);
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_macvlan_mode (const char *text, int state)
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
gs_free const char **words = NULL;
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
words = nm_utils_enum_get_values (nm_setting_macvlan_mode_get_type(),
NM_SETTING_MACVLAN_MODE_UNKNOWN + 1,
G_MAXINT);
return nmc_rl_gen_func_basic (text, state, words);
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
}
cli: fix some TAB completions for nmcli -a con add (cherry picked from commit 1cee85f556f7771e93c8637dac5ad284a405bf20)
2017-01-21 21:12:49 +01:00
static char *
gen_func_proxy_method (const char *text, int state)
{
gs_free const char **words = NULL;
words = nm_utils_enum_get_values (nm_setting_proxy_method_get_type (),
G_MININT,
G_MAXINT);
return nmc_rl_gen_func_basic (text, state, words);
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static char *
gen_func_master_ifnames (const char *text, int state)
cli: use a generic function asking for both IPv4/IPv6 in questionnaire
2013-09-19 13:54:04 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
int i;
GPtrArray *ifnames;
char *ret;
NMConnection *con;
NMSettingConnection *s_con;
const char *con_type, *ifname;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: use a generic function asking for both IPv4/IPv6 in questionnaire
2013-09-19 13:54:04 +02:00
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nm_cli.client);
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* Disable appending space after completion */
rl_completion_append_character = '\0';
cli: trivial: move is_property_valid() This removes one extra forward definition and saves one in future.
2015-06-07 15:26:43 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
ifnames = g_ptr_array_sized_new (20);
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
for (i = 0; i < connections->len; i++) {
con = NM_CONNECTION (connections->pdata[i]);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
s_con = nm_connection_get_setting_connection (con);
g_assert (s_con);
con_type = nm_setting_connection_get_connection_type (s_con);
if (g_strcmp0 (con_type, nmc_tab_completion.con_type) != 0)
continue;
ifname = nm_connection_get_interface_name (con);
g_ptr_array_add (ifnames, (gpointer) ifname);
}
g_ptr_array_add (ifnames, (gpointer) NULL);
cli: trivial: move is_property_valid() This removes one extra forward definition and saves one in future.
2015-06-07 15:26:43 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
ret = nmc_rl_gen_func_basic (text, state, (const char **) ifnames->pdata);
cli: trivial: move is_property_valid() This removes one extra forward definition and saves one in future.
2015-06-07 15:26:43 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
g_ptr_array_free (ifnames, TRUE);
cli: trivial: move is_property_valid() This removes one extra forward definition and saves one in future.
2015-06-07 15:26:43 +02:00
return ret;
}
cli: ask user before requiring optional parameters for 'nmcli -a con add type tun' Fixes: 64c6b124d24c0cd06b2868b4ca6a71f8e78e6c00
2015-11-27 22:32:31 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*----------------------------------------------------------------------------*/
nmcli: improve connection autocompletion (2/2) Complete the property as we parse the list of properties. This makes it possible to actually complete an unfinished property. E.g: $ nmcli --complete c modify enp0s25 +ipv6.addr +ipv6.addresses +ipv6.addr-gen-mode
2016-06-15 13:12:39 +02:00
cli: refactor: split connection property reader from do_connection_modify() No functional change. We'll need it for specifying the properties on connection addition.
2015-06-07 14:00:57 +02:00
static gboolean
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
set_connection_type (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
cli: refactor: split connection property reader from do_connection_modify() No functional change. We'll need it for specifying the properties on connection addition.
2015-06-07 14:00:57 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
const NameItem *type_settings, *slv_settings;
GError *local = NULL;
const gchar *master[] = { "master", NULL };
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
value = check_valid_name (value, nmc_valid_connection_types, NULL, &local);
if (!value) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: bad connection type: %s."), local->message);
g_clear_error (&local);
return FALSE;
}
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (g_strcmp0 (value, "bond-slave") == 0) {
value = NM_SETTING_WIRED_SETTING_NAME;
if (!set_property (con, NM_SETTING_CONNECTION_SETTING_NAME,
NM_SETTING_CONNECTION_SLAVE_TYPE, NM_SETTING_BOND_SETTING_NAME,
'\0', error)) {
return FALSE;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
enable_options (NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_MASTER, master);
} else if (g_strcmp0 (value, "bridge-slave") == 0) {
value = NM_SETTING_WIRED_SETTING_NAME;
if (!set_property (con, NM_SETTING_CONNECTION_SETTING_NAME,
NM_SETTING_CONNECTION_SLAVE_TYPE, NM_SETTING_BRIDGE_SETTING_NAME,
'\0', error)) {
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
return FALSE;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
}
enable_options (NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_MASTER, master);
} else if (g_strcmp0 (value, "team-slave") == 0) {
value = NM_SETTING_WIRED_SETTING_NAME;
if (!set_property (con, NM_SETTING_CONNECTION_SETTING_NAME,
NM_SETTING_CONNECTION_SLAVE_TYPE, NM_SETTING_TEAM_SETTING_NAME,
'\0', error)) {
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
return FALSE;
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
enable_options (NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_MASTER, master);
}
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* ifname is mandatory for all connection types except virtual ones (bond, team, bridge, vlan) */
if ( (strcmp (value, NM_SETTING_BOND_SETTING_NAME) == 0)
|| (strcmp (value, NM_SETTING_TEAM_SETTING_NAME) == 0)
|| (strcmp (value, NM_SETTING_BRIDGE_SETTING_NAME) == 0)
|| (strcmp (value, NM_SETTING_VLAN_SETTING_NAME) == 0)) {
disable_options (NM_SETTING_CONNECTION_SETTING_NAME,
NM_SETTING_CONNECTION_INTERFACE_NAME);
}
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!set_property (con, option->setting_name, option->property, value, '\0', error))
return FALSE;
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!con_settings (con, &type_settings, &slv_settings, error))
return FALSE;
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
ensure_settings (con, slv_settings);
ensure_settings (con, type_settings);
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return TRUE;
}
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_connection_iface (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
all: use nm_utils_is_valid_iface_name()
2016-12-23 12:51:26 +01:00
GError *tmp_error = NULL;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (value) {
/* Special value of '*' means no specific interface name */
if (strcmp (value, "*") == 0)
value = NULL;
all: use nm_utils_is_valid_iface_name()
2016-12-23 12:51:26 +01:00
else if (!nm_utils_is_valid_iface_name (value, &tmp_error)) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: '%s': %s"), value, tmp_error->message);
g_error_free (tmp_error);
return FALSE;
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
}
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return set_property (con, option->setting_name, option->property, value, '\0', error);
}
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_connection_master (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
NMSettingConnection *s_con;
const char *slave_type;
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
s_con = nm_connection_get_setting_connection (con);
g_return_val_if_fail (s_con, FALSE);
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!value) {
g_set_error_literal (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: master is required"));
return FALSE;
}
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
slave_type = nm_setting_connection_get_slave_type (s_con);
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
value = normalized_master_for_slave (connections, value, slave_type, &slave_type);
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!set_property (con, NM_SETTING_CONNECTION_SETTING_NAME,
NM_SETTING_CONNECTION_SLAVE_TYPE, slave_type,
'\0', error)) {
return FALSE;
}
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return set_property (con, option->setting_name, option->property, value, '\0', error);
}
cli: add macvlan support
2015-09-18 11:03:17 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_bond_option (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
NMSettingBond *s_bond;
gboolean success;
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
s_bond = nm_connection_get_setting_bond (con);
g_return_val_if_fail (s_bond, FALSE);
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!value)
return TRUE;
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (strcmp (option->option, "mode") == 0) {
value = nmc_bond_validate_mode (value, error);
if (!value)
return FALSE;
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (g_strcmp0 (value, "active-backup") == 0) {
const gchar *primary[] = { "primary", NULL };
enable_options (NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, primary);
}
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_MODE, value);
} else if (strcmp (option->option, "primary") == 0)
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_PRIMARY, value);
else if (strcmp (option->option, "miimon") == 0)
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_MIIMON, value);
else if (strcmp (option->option, "downdelay") == 0)
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_DOWNDELAY, value);
else if (strcmp (option->option, "updelay") == 0)
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_UPDELAY, value);
else if (strcmp (option->option, "arp-interval") == 0)
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_ARP_INTERVAL, value);
else if (strcmp (option->option, "arp-ip-target") == 0)
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_ARP_IP_TARGET, value);
else if (strcmp (option->option, "lacp-rate") == 0)
success = nm_setting_bond_add_option (s_bond, NM_SETTING_BOND_OPTION_LACP_RATE, value);
else
g_return_val_if_reached (FALSE);
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!success) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: error adding bond option '%s=%s'."),
option->option, value);
}
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return success;
}
cli: add support for tun settings Add nmcli support for adding and modifying TUN/TAP connections.
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_bond_monitoring_mode (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
NMSettingBond *s_bond;
gs_free gchar *monitor_mode = NULL;
const gchar *miimon_opts[] = { "miimon", "downdelay", "updelay", NULL };
const gchar *arp_opts[] = { "arp-interval", "arp-ip-target", NULL };
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
s_bond = nm_connection_get_setting_bond (con);
g_return_val_if_fail (s_bond, FALSE);
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (value) {
monitor_mode = g_strdup (value);
g_strstrip (monitor_mode);
} else {
monitor_mode = g_strdup (WORD_MIIMON);
}
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (monitor_mode, WORD_MIIMON))
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
enable_options (NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, miimon_opts);
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (monitor_mode, WORD_ARP))
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
enable_options (NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, arp_opts);
else {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: '%s' is not a valid monitoring mode; use '%s' or '%s'.\n"),
monitor_mode, WORD_MIIMON, WORD_ARP);
return FALSE;
}
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return TRUE;
}
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_bluetooth_type (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
NMSetting *setting;
cli: check ip-tunnel mode right after it is read
2015-12-16 16:12:24 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!value)
return TRUE;
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* 'dun' type requires adding 'gsm' or 'cdma' setting */
if ( !strcmp (value, NM_SETTING_BLUETOOTH_TYPE_DUN)
|| !strcmp (value, NM_SETTING_BLUETOOTH_TYPE_DUN"-gsm")) {
value = NM_SETTING_BLUETOOTH_TYPE_DUN;
setting = nm_setting_gsm_new ();
nmc_setting_custom_init (setting);
nm_connection_add_setting (con, setting);
} else if (!strcmp (value, NM_SETTING_BLUETOOTH_TYPE_DUN"-cdma")) {
value = NM_SETTING_BLUETOOTH_TYPE_DUN;
setting = nm_setting_cdma_new ();
nm_connection_add_setting (con, setting);
} else if (!strcmp (value, NM_SETTING_BLUETOOTH_TYPE_PANU)) {
/* no op */
} else {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: 'bt-type': '%s' not valid; use [%s, %s (%s), %s]."),
value, NM_SETTING_BLUETOOTH_TYPE_PANU, NM_SETTING_BLUETOOTH_TYPE_DUN,
NM_SETTING_BLUETOOTH_TYPE_DUN"-gsm", NM_SETTING_BLUETOOTH_TYPE_DUN"-cdma");
return FALSE;
}
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return set_property (con, option->setting_name, option->property, value, '\0', error);
}
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
set_yes_no (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
if (g_strcmp0 (value, _(WORD_LOC_YES)))
value = WORD_YES;
if (g_strcmp0 (value, _(WORD_LOC_NO)))
value = WORD_NO;
cli: allow specifying a 'dev' option when adding IP tunnels
2015-12-16 13:49:31 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return set_property (con, option->setting_name, option->property, value, '\0', error);
}
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
cli: default to method=manual when adding an address This restores accidentally changed behavior for "nmcli c add ... ip[46] <address>" Fixes: c5324ed285aff7d6d58212e4b030a0fc556eb43b (cherry picked from commit 195ad4ec976bc017b80bfbec80b2cd67d7b8dbdc)
2016-08-24 15:33:34 +02:00
static gboolean
set_ip4_address (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
NMSettingIPConfig *s_ip4;
if (!value)
return TRUE;
s_ip4 = nm_connection_get_setting_ip4_config (con);
if (!s_ip4) {
s_ip4 = (NMSettingIPConfig *) nm_setting_ip4_config_new ();
nm_connection_add_setting (con, NM_SETTING (s_ip4));
g_object_set (s_ip4,
NM_SETTING_IP_CONFIG_METHOD, NM_SETTING_IP4_CONFIG_METHOD_MANUAL,
NULL);
}
cli: properly set multiple addresses in questionnaire mode Pass the '+' modifier to set_property() for IPv4 and IPv6 addresses to append the new address to existing ones instead of overwriting them. Fixes: 2f456655593e92385a7e736691e7691514740805 https://bugzilla.redhat.com/show_bug.cgi?id=1380165
2016-09-28 22:39:09 +02:00
return set_property (con, option->setting_name, option->property, value,
option->flags & OPTION_MULTI ? '+' : '\0', error);
cli: default to method=manual when adding an address This restores accidentally changed behavior for "nmcli c add ... ip[46] <address>" Fixes: c5324ed285aff7d6d58212e4b030a0fc556eb43b (cherry picked from commit 195ad4ec976bc017b80bfbec80b2cd67d7b8dbdc)
2016-08-24 15:33:34 +02:00
}
static gboolean
set_ip6_address (NmCli *nmc, NMConnection *con, OptionInfo *option, const char *value, GError **error)
{
NMSettingIPConfig *s_ip6;
if (!value)
return TRUE;
s_ip6 = nm_connection_get_setting_ip6_config (con);
if (!s_ip6) {
s_ip6 = (NMSettingIPConfig *) nm_setting_ip6_config_new ();
nm_connection_add_setting (con, NM_SETTING (s_ip6));
g_object_set (s_ip6,
NM_SETTING_IP_CONFIG_METHOD, NM_SETTING_IP6_CONFIG_METHOD_MANUAL,
NULL);
}
cli: properly set multiple addresses in questionnaire mode Pass the '+' modifier to set_property() for IPv4 and IPv6 addresses to append the new address to existing ones instead of overwriting them. Fixes: 2f456655593e92385a7e736691e7691514740805 https://bugzilla.redhat.com/show_bug.cgi?id=1380165
2016-09-28 22:39:09 +02:00
return set_property (con, option->setting_name, option->property, value,
option->flags & OPTION_MULTI ? '+' : '\0', error);
cli: default to method=manual when adding an address This restores accidentally changed behavior for "nmcli c add ... ip[46] <address>" Fixes: c5324ed285aff7d6d58212e4b030a0fc556eb43b (cherry picked from commit 195ad4ec976bc017b80bfbec80b2cd67d7b8dbdc)
2016-08-24 15:33:34 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*----------------------------------------------------------------------------*/
cli: fix error handling in creation of IP tunnels Fixes: 7da440b798bd4f7d8f1344db5a1d3c98e57fcd5d
2015-12-16 11:11:06 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static OptionInfo option_info[] = {
{ NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_TYPE, "type", OPTION_REQD, PROMPT_CON_TYPE, NULL,
set_connection_type, gen_connection_types },
{ NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_ID, "con-name", OPTION_DONT_ASK, NULL, NULL, NULL, NULL },
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_AUTOCONNECT, "autoconnect", OPTION_DONT_ASK, NULL, NULL, NULL,
gen_func_bool_values_l10n },
cli: tab-complete "Interface name [*]" in questionnaire mode
2016-09-15 10:32:44 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_INTERFACE_NAME, "ifname", OPTION_REQD, PROMPT_IFNAME, NULL,
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
set_connection_iface, nmc_rl_gen_func_ifnames },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_MASTER, "master", OPTION_DONT_ASK, PROMPT_MASTER, NULL,
set_connection_master, gen_func_master_ifnames },
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
{ NM_SETTING_CONNECTION_SETTING_NAME, NM_SETTING_CONNECTION_SLAVE_TYPE, "slave-type", OPTION_DONT_ASK, NULL, NULL, NULL,
gen_func_slave_type },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_PPPOE_SETTING_NAME, NM_SETTING_PPPOE_USERNAME, "username", OPTION_REQD, N_("PPPoE username"), NULL, NULL, NULL },
{ NM_SETTING_PPPOE_SETTING_NAME, NM_SETTING_PPPOE_PASSWORD, "password", OPTION_NONE, N_("Password [none]"), NULL, NULL, NULL },
{ NM_SETTING_PPPOE_SETTING_NAME, NM_SETTING_PPPOE_SERVICE, "service", OPTION_NONE, N_("Service [none]"), NULL, NULL, NULL },
{ NM_SETTING_WIRED_SETTING_NAME, NM_SETTING_WIRED_MTU, "mtu", OPTION_NONE, N_("MTU [auto]"), NULL, NULL, NULL },
{ NM_SETTING_WIRED_SETTING_NAME, NM_SETTING_WIRED_MAC_ADDRESS, "mac", OPTION_NONE, N_("MAC [none]"), NULL, NULL, NULL },
{ NM_SETTING_WIRED_SETTING_NAME, NM_SETTING_WIRED_CLONED_MAC_ADDRESS, "cloned-mac", OPTION_NONE, N_("Cloned MAC [none]"), NULL, NULL, NULL },
{ NM_SETTING_INFINIBAND_SETTING_NAME, NM_SETTING_INFINIBAND_MTU, "mtu", OPTION_NONE, N_("MTU [auto]"), NULL, NULL, NULL },
{ NM_SETTING_INFINIBAND_SETTING_NAME, NM_SETTING_INFINIBAND_MAC_ADDRESS, "mac", OPTION_NONE, N_("MAC [none]"), NULL, NULL, NULL },
{ NM_SETTING_INFINIBAND_SETTING_NAME, NM_SETTING_INFINIBAND_TRANSPORT_MODE, "transport-mode", OPTION_NONE, PROMPT_IB_MODE, PROMPT_IB_MODE_CHOICES,
NULL, gen_func_ib_type },
{ NM_SETTING_INFINIBAND_SETTING_NAME, NM_SETTING_INFINIBAND_PARENT, "parent", OPTION_NONE, N_("Parent interface [none]"), NULL, NULL, NULL },
{ NM_SETTING_INFINIBAND_SETTING_NAME, NM_SETTING_INFINIBAND_P_KEY, "p-key", OPTION_NONE, N_("P_KEY [none]"), NULL, NULL, NULL },
{ NM_SETTING_WIRELESS_SETTING_NAME, NM_SETTING_WIRELESS_SSID, "ssid", OPTION_REQD, N_("SSID"), NULL, NULL, NULL },
{ NM_SETTING_WIRELESS_SETTING_NAME, NM_SETTING_WIRELESS_MODE, "mode", OPTION_NONE, PROMPT_WIFI_MODE, PROMPT_WIFI_MODE_CHOICES,
NULL, gen_func_wifi_mode },
{ NM_SETTING_WIRELESS_SETTING_NAME, NM_SETTING_WIRELESS_MTU, "mtu", OPTION_NONE, N_("MTU [auto]"), NULL, NULL, NULL },
{ NM_SETTING_WIRELESS_SETTING_NAME, NM_SETTING_WIRELESS_MAC_ADDRESS, "mac", OPTION_NONE, N_("MAC [none]"), NULL, NULL, NULL },
{ NM_SETTING_WIRELESS_SETTING_NAME, NM_SETTING_WIRELESS_CLONED_MAC_ADDRESS, "cloned-mac", OPTION_NONE, N_("Cloned MAC [none]"), NULL, NULL, NULL },
{ NM_SETTING_WIMAX_SETTING_NAME, NM_SETTING_WIMAX_NETWORK_NAME, "nsp", OPTION_REQD, N_("WiMAX NSP name"), NULL, NULL, NULL },
{ NM_SETTING_WIMAX_SETTING_NAME, NM_SETTING_WIMAX_MAC_ADDRESS, "mac", OPTION_NONE, N_("MAC [none]"), NULL, NULL, NULL },
{ NM_SETTING_GSM_SETTING_NAME, NM_SETTING_GSM_APN, "apn", OPTION_REQD, N_("APN"), NULL, NULL, NULL },
{ NM_SETTING_GSM_SETTING_NAME, NM_SETTING_GSM_USERNAME, "user", OPTION_NONE, N_("Username [none]"), NULL, NULL, NULL },
{ NM_SETTING_GSM_SETTING_NAME, NM_SETTING_GSM_PASSWORD, "password", OPTION_NONE, N_("Password [none]"), NULL, NULL, NULL },
{ NM_SETTING_CDMA_SETTING_NAME, NM_SETTING_CDMA_USERNAME, "user", OPTION_NONE, N_("Username [none]"), NULL, NULL, NULL },
{ NM_SETTING_CDMA_SETTING_NAME, NM_SETTING_CDMA_PASSWORD, "password", OPTION_NONE, N_("Password [none]"), NULL, NULL, NULL },
{ NM_SETTING_BLUETOOTH_SETTING_NAME, NM_SETTING_BLUETOOTH_BDADDR, "addr", OPTION_REQD, N_("Bluetooth device address"), NULL, NULL, NULL },
{ NM_SETTING_BLUETOOTH_SETTING_NAME, NM_SETTING_BLUETOOTH_TYPE, "bt-type", OPTION_NONE, PROMPT_BT_TYPE, PROMPT_BT_TYPE_CHOICES,
set_bluetooth_type, gen_func_bt_type },
{ NM_SETTING_VLAN_SETTING_NAME, NM_SETTING_VLAN_PARENT, "dev", OPTION_REQD, N_("VLAN parent device or connection UUID"), NULL,
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
NULL, nmc_rl_gen_func_ifnames },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_VLAN_SETTING_NAME, NM_SETTING_VLAN_ID, "id", OPTION_REQD, N_("VLAN ID (<0-4094>)"), NULL, NULL, NULL },
{ NM_SETTING_VLAN_SETTING_NAME, NM_SETTING_VLAN_FLAGS, "flags", OPTION_NONE, N_("VLAN flags (<0-7>) [none]"), NULL, NULL, NULL },
{ NM_SETTING_VLAN_SETTING_NAME, NM_SETTING_VLAN_INGRESS_PRIORITY_MAP, "ingress", OPTION_NONE, N_("Ingress priority maps [none]"), NULL, NULL, NULL },
{ NM_SETTING_VLAN_SETTING_NAME, NM_SETTING_VLAN_EGRESS_PRIORITY_MAP, "egress", OPTION_NONE, N_("Egress priority maps [none]"), NULL, NULL, NULL },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "mode", OPTION_NONE, PROMPT_BOND_MODE, "[balance-rr]",
set_bond_option, gen_func_bond_mode },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "primary", OPTION_DONT_ASK, N_("Bonding primary interface [none]"),
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
NULL, set_bond_option, nmc_rl_gen_func_ifnames },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, NULL, OPTION_NONE, N_("Bonding monitoring mode"), PROMPT_BOND_MON_MODE_CHOICES,
set_bond_monitoring_mode, gen_func_bond_mon_mode },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "miimon", OPTION_DONT_ASK, N_("Bonding miimon [100]"), NULL, set_bond_option, NULL },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "downdelay", OPTION_DONT_ASK, N_("Bonding downdelay [0]"), NULL, set_bond_option, NULL },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "updelay", OPTION_DONT_ASK, N_("Bonding updelay [0]"), NULL, set_bond_option, NULL },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "arp-interval", OPTION_DONT_ASK, N_("Bonding arp-interval [0]"), NULL,
set_bond_option, NULL },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "arp-ip-target", OPTION_DONT_ASK, N_("Bonding arp-ip-target [none]"),
NULL, set_bond_option, NULL },
{ NM_SETTING_BOND_SETTING_NAME, NM_SETTING_BOND_OPTIONS, "lacp-rate", OPTION_DONT_ASK, N_("LACP rate ('slow' or 'fast') [slow]"), NULL,
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
set_bond_option, gen_func_bond_lacp_rate },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_TEAM_SETTING_NAME, NM_SETTING_TEAM_CONFIG, "config", OPTION_NONE, N_("Team JSON configuration [none]"), NULL, NULL, NULL },
{ NM_SETTING_TEAM_PORT_SETTING_NAME, NM_SETTING_TEAM_PORT_CONFIG, "config", OPTION_NONE, N_("Team JSON configuration [none]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_STP, "stp", OPTION_NONE, N_("Enable STP [no]"), NULL,
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_PRIORITY, "priority", OPTION_NONE, N_("STP priority [32768]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_FORWARD_DELAY, "forward-delay", OPTION_NONE, N_("Forward delay [15]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_HELLO_TIME, "hello-time", OPTION_NONE, N_("Hello time [2]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_MAX_AGE, "max-age", OPTION_NONE, N_("Max age [20]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_AGEING_TIME, "ageing-time", OPTION_NONE, N_("MAC address ageing time [300]"), NULL, NULL, NULL },
cli: (trivial): remove stray %s
2016-09-15 00:12:37 +02:00
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_MULTICAST_SNOOPING, "multicast-snooping", OPTION_NONE, N_("Enable IGMP snooping [no]"), NULL,
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_BRIDGE_SETTING_NAME, NM_SETTING_BRIDGE_MAC_ADDRESS, "mac", OPTION_NONE, N_("MAC [none]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_PORT_SETTING_NAME, NM_SETTING_BRIDGE_PORT_PRIORITY, "priority", OPTION_NONE, N_("Bridge port priority [32]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_PORT_SETTING_NAME, NM_SETTING_BRIDGE_PORT_PATH_COST, "path-cost", OPTION_NONE, N_("Bridge port STP path cost [100]"), NULL, NULL, NULL },
{ NM_SETTING_BRIDGE_PORT_SETTING_NAME, NM_SETTING_BRIDGE_PORT_HAIRPIN_MODE, "hairpin", OPTION_NONE, N_("Hairpin [no]"), NULL,
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_VPN_SETTING_NAME, NM_SETTING_VPN_SERVICE_TYPE, "vpn-type", OPTION_REQD, PROMPT_VPN_TYPE, NULL, NULL, gen_func_vpn_types },
{ NM_SETTING_VPN_SETTING_NAME, NM_SETTING_VPN_USER_NAME, "user", OPTION_NONE, N_("Username [none]"), NULL, NULL, NULL },
{ NM_SETTING_OLPC_MESH_SETTING_NAME, NM_SETTING_OLPC_MESH_SSID, "ssid", OPTION_REQD, N_("SSID"), NULL, NULL, NULL },
{ NM_SETTING_OLPC_MESH_SETTING_NAME, NM_SETTING_OLPC_MESH_CHANNEL, "channel", OPTION_NONE, N_("OLPC Mesh channel [1]"), NULL, NULL, NULL },
{ NM_SETTING_OLPC_MESH_SETTING_NAME, NM_SETTING_OLPC_MESH_DHCP_ANYCAST_ADDRESS, "dhcp-anycast", OPTION_NONE, N_("DHCP anycast MAC address [none]"), NULL, NULL, NULL },
{ NM_SETTING_ADSL_SETTING_NAME, NM_SETTING_ADSL_USERNAME, "username", OPTION_REQD, N_("Username"), NULL, NULL, NULL },
{ NM_SETTING_ADSL_SETTING_NAME, NM_SETTING_ADSL_PROTOCOL, "protocol", OPTION_REQD, PROMPT_ADSL_PROTO, PROMPT_ADSL_PROTO_CHOICES,
NULL, gen_func_adsl_proto },
{ NM_SETTING_ADSL_SETTING_NAME, NM_SETTING_ADSL_PASSWORD, "password", OPTION_NONE, N_("Password [none]"), NULL, NULL, NULL },
{ NM_SETTING_ADSL_SETTING_NAME, NM_SETTING_ADSL_ENCAPSULATION, "encapsulation", OPTION_NONE, PROMPT_ADSL_ENCAP, PROMPT_ADSL_ENCAP_CHOICES,
NULL, gen_func_adsl_encap },
cli: macsec support
2016-07-01 11:56:33 +02:00
{ NM_SETTING_MACSEC_SETTING_NAME, NM_SETTING_MACSEC_PARENT, "dev", OPTION_REQD, N_("MACsec parent device or connection UUID"), NULL, NULL, NULL },
cli: fix some TAB completions for nmcli -a con add (cherry picked from commit 1cee85f556f7771e93c8637dac5ad284a405bf20)
2017-01-21 21:12:49 +01:00
{ NM_SETTING_MACSEC_SETTING_NAME, NM_SETTING_MACSEC_MODE, "mode", OPTION_REQD, PROMPT_MACSEC_MODE, PROMPT_MACSEC_MODE_CHOICES, NULL, gen_func_macsec_mode },
cli: macsec support
2016-07-01 11:56:33 +02:00
{ NM_SETTING_MACSEC_SETTING_NAME, NM_SETTING_MACSEC_ENCRYPT, "encrypt", OPTION_NONE, N_("Enable encryption [yes]"), NULL, set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_MACSEC_SETTING_NAME, NM_SETTING_MACSEC_MKA_CAK, "cak", OPTION_NONE, N_("MKA CAK"), NULL, NULL, NULL },
{ NM_SETTING_MACSEC_SETTING_NAME, NM_SETTING_MACSEC_MKA_CKN, "ckn", OPTION_NONE, N_("MKA_CKN"), NULL, NULL, NULL },
{ NM_SETTING_MACSEC_SETTING_NAME, NM_SETTING_MACSEC_PORT, "port", OPTION_NONE, N_("SCI port [1]"), NULL, NULL, NULL },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_MACVLAN_SETTING_NAME, NM_SETTING_MACVLAN_PARENT, "dev", OPTION_REQD, N_("MACVLAN parent device or connection UUID"), NULL,
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
NULL, nmc_rl_gen_func_ifnames },
{ NM_SETTING_MACVLAN_SETTING_NAME, NM_SETTING_MACVLAN_MODE, "mode", OPTION_REQD, PROMPT_MACVLAN_MODE, NULL,
NULL, gen_func_macvlan_mode },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_MACVLAN_SETTING_NAME, NM_SETTING_MACVLAN_TAP, "tap", OPTION_NONE, N_("Tap [no]"), NULL,
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_VXLAN_SETTING_NAME, NM_SETTING_VXLAN_ID, "id", OPTION_REQD, N_("VXLAN ID"), NULL, NULL, NULL },
{ NM_SETTING_VXLAN_SETTING_NAME, NM_SETTING_VXLAN_REMOTE, "remote", OPTION_REQD, N_("Remote"), NULL, NULL, NULL },
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
{ NM_SETTING_VXLAN_SETTING_NAME, NM_SETTING_VXLAN_PARENT, "dev", OPTION_NONE, N_("Parent device [none]"), NULL,
NULL, nmc_rl_gen_func_ifnames },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_VXLAN_SETTING_NAME, NM_SETTING_VXLAN_LOCAL, "local", OPTION_NONE, N_("Local address [none]"), NULL, NULL, NULL },
{ NM_SETTING_VXLAN_SETTING_NAME, NM_SETTING_VXLAN_SOURCE_PORT_MIN, "source-port-min", OPTION_NONE, N_("Minimum source port [0]"), NULL, NULL, NULL },
{ NM_SETTING_VXLAN_SETTING_NAME, NM_SETTING_VXLAN_SOURCE_PORT_MAX, "source-port-max", OPTION_NONE, N_("Maximum source port [0]"), NULL, NULL, NULL },
{ NM_SETTING_VXLAN_SETTING_NAME, NM_SETTING_VXLAN_DESTINATION_PORT, "destination-port", OPTION_NONE, N_("Destination port [8472]"), NULL, NULL, NULL },
{ NM_SETTING_TUN_SETTING_NAME, NM_SETTING_TUN_MODE, "mode", OPTION_NONE, PROMPT_TUN_MODE, PROMPT_TUN_MODE_CHOICES,
NULL, gen_func_tun_mode },
{ NM_SETTING_TUN_SETTING_NAME, NM_SETTING_TUN_OWNER, "owner", OPTION_NONE, N_("User ID [none]"), NULL, NULL, NULL },
{ NM_SETTING_TUN_SETTING_NAME, NM_SETTING_TUN_GROUP, "group", OPTION_NONE, N_("Group ID [none]"), NULL, NULL, NULL },
{ NM_SETTING_TUN_SETTING_NAME, NM_SETTING_TUN_PI, "pi", OPTION_NONE, N_("Enable PI [no]"), NULL,
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_TUN_SETTING_NAME, NM_SETTING_TUN_VNET_HDR, "vnet-hdr", OPTION_NONE, N_("Enable VNET header [no]"), NULL,
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_TUN_SETTING_NAME, NM_SETTING_TUN_MULTI_QUEUE, "multi-queue", OPTION_NONE, N_("Enable multi queue [no]"), NULL,
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_IP_TUNNEL_SETTING_NAME, NM_SETTING_IP_TUNNEL_MODE, "mode", OPTION_REQD, PROMPT_IP_TUNNEL_MODE, NULL, NULL, gen_func_ip_tunnel_mode },
{ NM_SETTING_IP_TUNNEL_SETTING_NAME, NM_SETTING_IP_TUNNEL_LOCAL, "local", OPTION_NONE, N_("Local endpoint [none]"), NULL, NULL, NULL },
{ NM_SETTING_IP_TUNNEL_SETTING_NAME, NM_SETTING_IP_TUNNEL_REMOTE, "remote", OPTION_REQD, N_("Remote"), NULL, NULL, NULL },
cli: fix completion/add missing functions for '--complete-args' (rh #1375933) This makes bash completion work again for 'nmcli connection add'. Fixes: 8b39090597159dea52b5ffb760395ab593c7a3dc https://bugzilla.redhat.com/show_bug.cgi?id=1375933
2016-09-15 00:11:17 +02:00
{ NM_SETTING_IP_TUNNEL_SETTING_NAME, NM_SETTING_IP_TUNNEL_PARENT, "dev", OPTION_NONE, N_("Parent device [none]"), NULL,
NULL, nmc_rl_gen_func_ifnames },
cli: default to method=manual when adding an address This restores accidentally changed behavior for "nmcli c add ... ip[46] <address>" Fixes: c5324ed285aff7d6d58212e4b030a0fc556eb43b (cherry picked from commit 195ad4ec976bc017b80bfbec80b2cd67d7b8dbdc)
2016-08-24 15:33:34 +02:00
{ NM_SETTING_IP4_CONFIG_SETTING_NAME, NM_SETTING_IP_CONFIG_ADDRESSES, "ip4", OPTION_MULTI, N_("IPv4 address (IP[/plen]) [none]"), NULL,
set_ip4_address, NULL },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_IP4_CONFIG_SETTING_NAME, NM_SETTING_IP_CONFIG_GATEWAY, "gw4", OPTION_NONE, N_("IPv4 gateway [none]"), NULL, NULL, NULL },
cli: default to method=manual when adding an address This restores accidentally changed behavior for "nmcli c add ... ip[46] <address>" Fixes: c5324ed285aff7d6d58212e4b030a0fc556eb43b (cherry picked from commit 195ad4ec976bc017b80bfbec80b2cd67d7b8dbdc)
2016-08-24 15:33:34 +02:00
{ NM_SETTING_IP6_CONFIG_SETTING_NAME, NM_SETTING_IP_CONFIG_ADDRESSES, "ip6", OPTION_MULTI, N_("IPv6 address (IP[/plen]) [none]"), NULL,
set_ip6_address, NULL },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NM_SETTING_IP6_CONFIG_SETTING_NAME, NM_SETTING_IP_CONFIG_GATEWAY, "gw6", OPTION_NONE, N_("IPv6 gateway [none]"), NULL, NULL, NULL },
cli: fix some TAB completions for nmcli -a con add (cherry picked from commit 1cee85f556f7771e93c8637dac5ad284a405bf20)
2017-01-21 21:12:49 +01:00
{ NM_SETTING_PROXY_SETTING_NAME, NM_SETTING_PROXY_METHOD, "method", OPTION_NONE, PROMPT_PROXY_METHOD, PROMPT_PROXY_METHOD_CHOICES, NULL, gen_func_proxy_method },
{ NM_SETTING_PROXY_SETTING_NAME, NM_SETTING_PROXY_BROWSER_ONLY, "browser-only", OPTION_NONE, N_("Browser only [no]"), NULL,
set_yes_no, gen_func_bool_values_l10n },
{ NM_SETTING_PROXY_SETTING_NAME, NM_SETTING_PROXY_PAC_URL, "pac-url", OPTION_NONE, N_("PAC URL"), NULL, NULL, NULL },
{ NM_SETTING_PROXY_SETTING_NAME, NM_SETTING_PROXY_PAC_SCRIPT, "pac-script", OPTION_NONE, N_("PAC script"), NULL, NULL, NULL },
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{ NULL, NULL, NULL, OPTION_NONE, NULL, NULL, NULL, NULL },
};
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
option_relevant (NMConnection *connection, OptionInfo *option)
{
if (option->flags & OPTION_DONT_ASK && !(option->flags & OPTION_ENABLED))
return FALSE;
if (option->flags & OPTION_DISABLED)
return FALSE;
if (!nm_connection_get_setting_by_name (connection, option->setting_name))
return FALSE;
return TRUE;
}
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*----------------------------------------------------------------------------*/
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
complete_property_name (NmCli *nmc, NMConnection *connection,
char modifier,
const gchar *prefix,
const gchar *postfix)
{
NMSettingConnection *s_con;
const NameItem *valid_settings_main = NULL;
const NameItem *valid_settings_slave = NULL;
const char *connection_type = NULL;
const char *slave_type = NULL;
gs_free char *slv_type = NULL;
gs_free char *word_list = NULL;
OptionInfo *candidate;
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
connection_type = nm_connection_get_connection_type (connection);
s_con = nm_connection_get_setting_connection (connection);
if (s_con)
slave_type = nm_setting_connection_get_slave_type (s_con);
slv_type = g_strdup_printf ("%s-slave", slave_type ? slave_type : "no");
valid_settings_main = get_valid_settings_array (connection_type);
valid_settings_slave = get_valid_settings_array (slv_type);
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
word_list = get_valid_properties_string (valid_settings_main, valid_settings_slave, modifier, prefix, postfix);
if (word_list)
g_print ("%s", word_list);
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (modifier != '\0')
return;
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
for (candidate = option_info; candidate->setting_name; candidate++) {
if (!nm_connection_get_setting_by_name (connection, candidate->setting_name))
continue;
if (!candidate->option)
continue;
if (!g_str_has_prefix (candidate->option, prefix))
continue;
man: update the nmcli manual for new connection add syntax It allows us to clean up the nmcli "c add" section considerably. We list the old-fashioned aliases in a separate section that applies to both "nmcli c add" and "nmcli c modify". The section is now nicely cross-linked with nm-settings in HTML rendering.
2016-05-27 21:06:14 +02:00
g_print ("%s\n", candidate->option);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
}
}
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
run_rl_generator (rl_compentry_func_t *generator_func, const char *prefix)
{
int state = 0;
char *str;
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
while ((str = generator_func (prefix, state))) {
man: update the nmcli manual for new connection add syntax It allows us to clean up the nmcli "c add" section considerably. We list the old-fashioned aliases in a separate section that applies to both "nmcli c add" and "nmcli c modify". The section is now nicely cross-linked with nm-settings in HTML rendering.
2016-05-27 21:06:14 +02:00
g_print ("%s\n", str);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
g_free (str);
if (state == 0)
state = 1;
}
}
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
complete_option (OptionInfo *option, const gchar *prefix)
{
if (option->generator_func)
run_rl_generator (option->generator_func, prefix);
}
cli: add vxlan support
2015-10-13 18:53:10 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
complete_property (const gchar *setting_name, const gchar *property, const gchar *prefix)
{
if (strcmp (setting_name, NM_SETTING_CONNECTION_SETTING_NAME) == 0) {
if (strcmp (property, NM_SETTING_CONNECTION_TYPE) == 0)
run_rl_generator (gen_connection_types, prefix);
else if (strcmp (property, NM_SETTING_CONNECTION_MASTER) == 0)
run_rl_generator (gen_func_master_ifnames, prefix);
cli: enable bash completion for some more properties of add/modify connection.interface-name mavclan.tap mavclan.parent ip-tunnel.parent vxlan.parent
2016-09-15 02:49:21 +02:00
else if (strcmp (property, NM_SETTING_CONNECTION_INTERFACE_NAME) == 0)
run_rl_generator (nmc_rl_gen_func_ifnames, prefix);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
} else if ( strcmp (setting_name, NM_SETTING_VPN_SETTING_NAME) == 0
&& strcmp (property, NM_SETTING_VPN_SERVICE_TYPE) == 0)
run_rl_generator (gen_func_vpn_types, prefix);
else if ( strcmp (setting_name, NM_SETTING_WIRELESS_SETTING_NAME) == 0
&& strcmp (property, NM_SETTING_WIRELESS_MODE) == 0)
run_rl_generator (gen_func_wifi_mode, prefix);
else if ( strcmp (setting_name, NM_SETTING_INFINIBAND_SETTING_NAME) == 0
&& strcmp (property, NM_SETTING_INFINIBAND_TRANSPORT_MODE) == 0)
run_rl_generator (gen_func_ib_type, prefix);
else if ( strcmp (setting_name, NM_SETTING_BLUETOOTH_SETTING_NAME) == 0
&& strcmp (property, NM_SETTING_BLUETOOTH_TYPE) == 0)
run_rl_generator (gen_func_bt_type, prefix);
else if (strcmp (setting_name, NM_SETTING_ADSL_SETTING_NAME) == 0) {
if (strcmp (property, NM_SETTING_ADSL_PROTOCOL) == 0)
run_rl_generator (gen_func_adsl_proto, prefix);
else if (strcmp (property, NM_SETTING_ADSL_ENCAPSULATION) == 0)
run_rl_generator (gen_func_adsl_encap, prefix);
} else if ( strcmp (setting_name, NM_SETTING_TUN_SETTING_NAME) == 0
&& strcmp (property, NM_SETTING_TUN_MODE) == 0)
run_rl_generator (gen_func_tun_mode, prefix);
cli: enable bash completion for some more properties of add/modify connection.interface-name mavclan.tap mavclan.parent ip-tunnel.parent vxlan.parent
2016-09-15 02:49:21 +02:00
else if (strcmp (setting_name, NM_SETTING_IP_TUNNEL_SETTING_NAME) == 0) {
if (strcmp (property, NM_SETTING_IP_TUNNEL_MODE) == 0)
run_rl_generator (gen_func_ip_tunnel_mode, prefix);
else if (strcmp (property, NM_SETTING_IP_TUNNEL_PARENT) == 0)
run_rl_generator (nmc_rl_gen_func_ifnames, prefix);
} else if (strcmp (setting_name, NM_SETTING_MACVLAN_SETTING_NAME) == 0) {
if (strcmp (property, NM_SETTING_MACVLAN_MODE) == 0)
run_rl_generator (gen_func_macvlan_mode, prefix);
else if (strcmp (property, NM_SETTING_MACVLAN_PARENT) == 0)
run_rl_generator (nmc_rl_gen_func_ifnames, prefix);
else if (strcmp (property, NM_SETTING_MACVLAN_TAP) == 0)
run_rl_generator (gen_func_bool_values_l10n, prefix);
} else if ( strcmp (setting_name, NM_SETTING_VXLAN_SETTING_NAME) == 0
&& strcmp (property, NM_SETTING_VXLAN_PARENT) == 0)
run_rl_generator (nmc_rl_gen_func_ifnames, prefix);
cli: tab-complete "Interface name [*]" in questionnaire mode
2016-09-15 10:32:44 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
}
cli: add support for IP tunnel settings
2015-09-01 15:14:03 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/*----------------------------------------------------------------------------*/
static gboolean
get_value (const char **value, int *argc, char ***argv, const char *option, GError **error)
{
if (!**argv) {
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
_("Error: value for '%s' is missing."), option);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
return FALSE;
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* Empty string will reset the value to default */
if (**argv[0] == '\0')
*value = NULL;
else
*value = *argv[0];
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
(*argc)--;
(*argv)++;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return TRUE;
}
cli: 'connection add': accept multiple IP addresses
2012-12-04 15:08:14 +01:00
cli/connection: export read_connection_properties() It makes sense to modify the applied connection from the device object.
2016-06-22 17:35:16 +02:00
gboolean
nmc_read_connection_properties (NmCli *nmc,
NMConnection *connection,
int *argc,
char ***argv,
GError **error)
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
{
const char *option;
const char *value = NULL;
GError *local = NULL;
cli: 'connection add': accept multiple IP addresses
2012-12-04 15:08:14 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* First check if we have a slave-type, as this would mean we will not
* have ip properties but possibly others, slave-type specific.
*/
/* Go through arguments and set properties */
do {
OptionInfo *candidate;
OptionInfo *chosen = NULL;
gs_strfreev gchar **strv = NULL;
const NameItem *type_settings, *slv_settings;
char modifier = '\0';
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!con_settings (connection, &type_settings, &slv_settings, error))
return FALSE;
cli: 'connection add': accept multiple IP addresses
2012-12-04 15:08:14 +01:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
ensure_settings (connection, slv_settings);
ensure_settings (connection, type_settings);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
option = **argv;
if (!option) {
g_set_error_literal (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: <setting>.<property> argument is missing."));
return FALSE;
}
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (option[0] == '+' || option[0] == '-')
modifier = *option;
strv = g_strsplit (option, ".", 2);
if (g_strv_length (strv) == 2) {
/* This seems like a <setting>.<property> (such as "connection.id" or "bond.mode"),
* optionally prefixed with "+| or "-". */
char *setting = strv[0];
const char *setting_name;
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (modifier)
setting++;
cli/connections: only do completion for the last argument Completing the property when we stop parsing due to error is not the right thing to do.
2016-06-23 19:59:40 +02:00
if (*argc == 1 && nmc->complete)
complete_property_name (nmc, connection, modifier, setting, strv[1]);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
setting_name = check_valid_name (setting, type_settings, slv_settings, &local);
if (!setting_name) {
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: invalid or not allowed setting '%s': %s."),
setting, local->message);
g_clear_error (&local);
return FALSE;
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
}
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
(*argc)--;
(*argv)++;
cli/connections: only do completion for the last argument Completing the property when we stop parsing due to error is not the right thing to do.
2016-06-23 19:59:40 +02:00
if (!get_value (&value, argc, argv, option, error))
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return FALSE;
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!*argc && nmc->complete)
complete_property (setting, strv[1], value ? value : "");
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!set_property (connection, setting_name, strv[1], value, modifier, error))
return FALSE;
} else {
/* Let's see if this is an property alias (such as "id", "mode", "type" or "con-name")*/
for (candidate = option_info; candidate->setting_name; candidate++) {
if (g_strcmp0 (candidate->option, option))
continue;
if (!check_valid_name (candidate->setting_name, type_settings, slv_settings, NULL))
continue;
if (chosen) {
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
_("Error: '%s' is ambiguous (%s.%s or %s.%s)."), option,
chosen->setting_name, chosen->property,
candidate->setting_name, candidate->property);
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
return FALSE;
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
chosen = candidate;
}
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!chosen) {
if (modifier)
option++;
cli/connections: only do completion for the last argument Completing the property when we stop parsing due to error is not the right thing to do.
2016-06-23 19:59:40 +02:00
if (*argc == 1 && nmc->complete)
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
complete_property_name (nmc, connection, modifier, option, NULL);
g_set_error (error, NMCLI_ERROR, NMC_RESULT_ERROR_USER_INPUT,
_("Error: invalid <setting>.<property> '%s'."), option);
return FALSE;
libnm-core: add NMSettingIPConfig:gateway, drop NMIPAddress:gateway The gateway is a global property of the IPv4/IPv6 configuration, not an attribute of any particular address. So represent it as such in the API; remove the gateway from NMIPAddress, and add it to NMSettingIPConfig. Behind the scenes, the gateway is still serialized along with the first address in NMSettingIPConfig:addresses, and is deserialized from that if the settings dictionary doesn't contain a 'gateway' key. Adjust nmcli's interactive mode to prompt for IP addresses and gateway separately. (Patch partly from Jirka Klimeš.)
2014-10-20 21:30:56 -04:00
}
cli: 'connection add' - questionnaire mode (rh #953291) Synopsis: nmcli --ask connection add When '--ask' is used, nmcli will ask not only for mandatory arguments but also for the optional ones, with a series of questions.
2013-08-27 13:13:23 +01:00
cli/connections: only do completion for the last argument Completing the property when we stop parsing due to error is not the right thing to do.
2016-06-23 19:59:40 +02:00
if (*argc == 1 && nmc->complete)
complete_property_name (nmc, connection, modifier, option, NULL);
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
(*argc)--;
(*argv)++;
cli/connections: only do completion for the last argument Completing the property when we stop parsing due to error is not the right thing to do.
2016-06-23 19:59:40 +02:00
if (!get_value (&value, argc, argv, option, error))
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
return FALSE;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!*argc && nmc->complete)
complete_option (chosen, value ? value : "");
cli: allow specifying arbitrary properties on "nmcli c add" Syntax: nmcli c add ... -- [+|-]<setting>.<property> <value> ...
2015-06-07 15:35:32 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!set_option (nmc, connection, chosen, value, error))
return FALSE;
}
cli: allow specifying arbitrary properties on "nmcli c add" Syntax: nmcli c add ... -- [+|-]<setting>.<property> <value> ...
2015-06-07 15:35:32 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
} while (*argc);
cli: allow specifying arbitrary properties on "nmcli c add" Syntax: nmcli c add ... -- [+|-]<setting>.<property> <value> ...
2015-06-07 15:35:32 +02:00
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
return TRUE;
}
typedef struct {
NmCli *nmc;
char *con_name;
} AddConnectionInfo;
static void
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
add_connection_cb (GObject *client,
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
GAsyncResult *result,
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
gpointer user_data)
{
AddConnectionInfo *info = (AddConnectionInfo *) user_data;
NmCli *nmc = info->nmc;
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
NMRemoteConnection *connection;
GError *error = NULL;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
connection = nm_client_add_connection_finish (NM_CLIENT (client), result, &error);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
if (error) {
g_string_printf (nmc->return_text,
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
_("Error: Failed to add '%s' connection: %s"),
info->con_name, error->message);
g_error_free (error);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmc->return_value = NMC_RESULT_ERROR_CON_ACTIVATION;
} else {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Connection '%s' (%s) successfully added.\n"),
nm_connection_get_id (NM_CONNECTION (connection)),
nm_connection_get_uuid (NM_CONNECTION (connection)));
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
g_object_unref (connection);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
g_free (info->con_name);
g_free (info);
quit ();
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
add_new_connection (gboolean persistent,
NMClient *client,
NMConnection *connection,
GAsyncReadyCallback callback,
gpointer user_data)
cli: enable TAB-completion for "MACVLAN mode: " in nmcli -a con add type macvlan
2015-12-16 16:09:18 +01:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
nm_client_add_connection_async (client, connection, persistent,
NULL, callback, user_data);
cli: enable TAB-completion for "MACVLAN mode: " in nmcli -a con add type macvlan
2015-12-16 16:09:18 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
update_connection (gboolean persistent,
NMRemoteConnection *connection,
GAsyncReadyCallback callback,
gpointer user_data)
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
nm_remote_connection_commit_changes_async (connection, persistent,
NULL, callback, user_data);
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
}
static gboolean
is_single_word (const char* line)
{
size_t n1, n2, n3;
n1 = strspn (line, " \t");
n2 = strcspn (line+n1, " \t\0") + n1;
n3 = strspn (line+n2, " \t");
if (n3 == 0)
return TRUE;
else
return FALSE;
}
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
static char **
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
nmcli_con_add_tab_completion (const char *text, int start, int end)
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
{
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
char **match_array = NULL;
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
rl_compentry_func_t *generator_func = NULL;
cli: fix yes/no completion in questionnaire mode
2016-09-15 03:05:00 +02:00
gs_free char *no = g_strdup_printf ("[%s]: ", gettext ("no"));
gs_free char *yes = g_strdup_printf ("[%s]: ", gettext ("yes"));
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
/* Disable readline's default filename completion */
rl_attempted_completion_over = 1;
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
/* Restore standard append character to space */
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
rl_completion_append_character = '\x00';
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
if (!is_single_word (rl_line_buffer))
return NULL;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (g_str_has_prefix (rl_prompt, PROMPT_CON_TYPE))
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
generator_func = gen_connection_types;
cli: tab-complete "Interface name [*]" in questionnaire mode
2016-09-15 10:32:44 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_IFNAME))
generator_func = nmc_rl_gen_func_ifnames;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_VPN_TYPE))
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
generator_func = gen_func_vpn_types;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_MASTER))
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
generator_func = gen_func_master_ifnames;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_WIFI_MODE))
cli: Add mode option for wifi connecitons Only default (infrastructure) mode connections can be created and as it's not possible to write mode=ap connections with ifcfg-rh plugins, they can't be switched to mode=ap.
2014-10-04 17:07:42 +02:00
generator_func = gen_func_wifi_mode;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_IB_MODE))
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
generator_func = gen_func_ib_type;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_BT_TYPE))
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
generator_func = gen_func_bt_type;
else if (g_str_has_prefix (rl_prompt, PROMPT_BOND_MODE))
generator_func = gen_func_bond_mode;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_BOND_MON_MODE))
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
generator_func = gen_func_bond_mon_mode;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_ADSL_PROTO))
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
generator_func = gen_func_adsl_proto;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_ADSL_ENCAP))
cli: allow creating ADSL connections with 'nmcli connection add' (rh #1264089) https://bugzilla.redhat.com/show_bug.cgi?id=1264089
2015-09-18 12:05:49 +02:00
generator_func = gen_func_adsl_encap;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_TUN_MODE))
cli: allow TAB-completion of "Mode:" in 'nmcli -a con add type tun' Fixes: 64c6b124d24c0cd06b2868b4ca6a71f8e78e6c00
2015-11-27 23:10:12 +01:00
generator_func = gen_func_tun_mode;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_IP_TUNNEL_MODE))
cli: enable TAB-completion for "Tunnel mode: " in nmcli -a con add type ip-tun
2015-12-16 14:36:28 +01:00
generator_func = gen_func_ip_tunnel_mode;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if (g_str_has_prefix (rl_prompt, PROMPT_MACVLAN_MODE))
cli: enable TAB-completion for "MACVLAN mode: " in nmcli -a con add type macvlan
2015-12-16 16:09:18 +01:00
generator_func = gen_func_macvlan_mode;
cli: fix some TAB completions for nmcli -a con add (cherry picked from commit 1cee85f556f7771e93c8637dac5ad284a405bf20)
2017-01-21 21:12:49 +01:00
else if (g_str_has_prefix (rl_prompt, PROMPT_MACSEC_MODE))
generator_func = gen_func_macsec_mode;
else if (g_str_has_prefix (rl_prompt, PROMPT_PROXY_METHOD))
generator_func = gen_func_proxy_method;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
else if ( g_str_has_suffix (rl_prompt, yes)
|| g_str_has_suffix (rl_prompt, no))
generator_func = gen_func_bool_values_l10n;
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
if (generator_func)
match_array = rl_completion_matches (text, generator_func);
return match_array;
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
ask_option (NmCli *nmc, NMConnection *connection, OptionInfo *option)
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
{
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
gchar *value;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
GError *error = NULL;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
gs_free gchar *prompt = NULL;
gboolean multi = option->flags & OPTION_MULTI;
prompt = g_strjoin ("",
gettext (option->prompt),
option->def_hint ? " " : "",
option->def_hint ? option->def_hint : "",
": ",
NULL);
if (multi)
g_print (_("You can specify this option more than once. Press <Enter> when you're done.\n"));
again:
value = nmc_readline ("%s", prompt);
if (multi && !value)
return;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!set_option (nmc, connection, option, value, &error)) {
g_printerr ("%s\n", error->message);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
g_clear_error (&error);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
goto again;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (multi && value)
goto again;
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static void
questionnaire_mandatory (NmCli *nmc, NMConnection *connection)
{
OptionInfo *candidate;
/* Mandatory settings. */
for (candidate = option_info; candidate->setting_name; candidate++) {
if (!option_relevant (connection, candidate))
continue;
if (candidate->flags & OPTION_REQD || candidate->flags & OPTION_ENABLED)
ask_option (nmc, connection, candidate);
cli: allow creating temporary (unsaved) connection with 'nmcli con add'
2014-01-24 11:04:17 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
}
cli: 'con add': make ifname mandatory (except bond,bridge,vlan) (bgo #698113) Optional 'ifname' allowed creating connection applicable to all interfaces, which was confusing for some users. Now we require the user to provide ifname to lock the connection for an interface. An "unbound" connection can be created with ifname "*". $ nmcli connection add type eth ifname eth0 $ nmcli connection add type eth now becomes $ nmcli connection add type eth ifname "*" bond, bridge: - when ifname is not specified or is "*", interface name is generated (nm-bond, nm-bridge) vlan: - when ifname is not specified or is "*", vlan device is named "dev.id" Note: the quotes around * are required to suppress shell expansion.
2013-06-10 15:31:09 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
static gboolean
want_provide_opt_args (const char *type, int num)
{
char *answer;
gboolean ret = TRUE;
cli: 'con add': make ifname mandatory (except bond,bridge,vlan) (bgo #698113) Optional 'ifname' allowed creating connection applicable to all interfaces, which was confusing for some users. Now we require the user to provide ifname to lock the connection for an interface. An "unbound" connection can be created with ifname "*". $ nmcli connection add type eth ifname eth0 $ nmcli connection add type eth now becomes $ nmcli connection add type eth ifname "*" bond, bridge: - when ifname is not specified or is "*", interface name is generated (nm-bond, nm-bridge) vlan: - when ifname is not specified or is "*", vlan device is named "dev.id" Note: the quotes around * are required to suppress shell expansion.
2013-06-10 15:31:09 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* Ask for optional arguments. */
g_print (ngettext ("There is %d optional setting for %s.\n",
"There are %d optional settings for %s.\n", num),
num, type);
answer = nmc_readline (ngettext ("Do you want to provide it? %s",
"Do you want to provide them? %s", num),
prompt_yes_no (TRUE, NULL));
answer = answer ? g_strstrip (answer) : NULL;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (answer && !matches (answer, WORD_LOC_YES))
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
ret = FALSE;
g_free (answer);
return ret;
}
static const char *
setting_name_to_name (const char *name)
{
if (strcmp (name, NM_SETTING_WIRED_SETTING_NAME) == 0)
return _("Wired Ethernet");
if (strcmp (name, NM_SETTING_INFINIBAND_SETTING_NAME) == 0)
return _("InfiniBand connection");
if (strcmp (name, NM_SETTING_WIRELESS_SETTING_NAME) == 0)
return _("Wi-Fi connection");
if (strcmp (name, NM_SETTING_WIMAX_SETTING_NAME) == 0)
return _("WiMAX connection");
if (strcmp (name, NM_SETTING_PPPOE_SETTING_NAME) == 0)
return _("PPPoE");
if (strcmp (name, NM_SETTING_CDMA_SETTING_NAME) == 0)
return _("CDMA mobile broadband connection");
if (strcmp (name, NM_SETTING_GSM_SETTING_NAME) == 0)
return _("GSM mobile broadband connection");
if (strcmp (name, NM_SETTING_BLUETOOTH_SETTING_NAME) == 0)
return _("bluetooth connection");
if (strcmp (name, NM_SETTING_VLAN_SETTING_NAME) == 0)
return _("VLAN connection");
if (strcmp (name, NM_SETTING_BOND_SETTING_NAME) == 0)
return _("Bond device");
if (strcmp (name, NM_SETTING_TEAM_SETTING_NAME) == 0)
return _("Team device");
if (strcmp (name, NM_SETTING_TEAM_PORT_SETTING_NAME) == 0)
return _("Team port");
if (strcmp (name, NM_SETTING_BRIDGE_SETTING_NAME) == 0)
return _("Bridge device");
if (strcmp (name, NM_SETTING_BRIDGE_PORT_SETTING_NAME) == 0)
return _("Bridge port");
if (strcmp (name, NM_SETTING_VPN_SETTING_NAME) == 0)
return _("VPN connection");
if (strcmp (name, NM_SETTING_OLPC_MESH_SETTING_NAME) == 0)
return _("OLPC Mesh connection");
if (strcmp (name, NM_SETTING_ADSL_SETTING_NAME) == 0)
return _("ADSL connection");
cli: macsec support
2016-07-01 11:56:33 +02:00
if (strcmp (name, NM_SETTING_MACSEC_SETTING_NAME) == 0)
return _("MACsec connection");
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (strcmp (name, NM_SETTING_MACVLAN_SETTING_NAME) == 0)
return _("macvlan connection");
if (strcmp (name, NM_SETTING_VXLAN_SETTING_NAME) == 0)
return _("VXLAN connection");
if (strcmp (name, NM_SETTING_TUN_SETTING_NAME) == 0)
return _("Tun device");
if (strcmp (name, NM_SETTING_IP4_CONFIG_SETTING_NAME) == 0)
return _("IPv4 protocol");
if (strcmp (name, NM_SETTING_IP6_CONFIG_SETTING_NAME) == 0)
return _("IPv6 protocol");
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
if (strcmp (name, NM_SETTING_PROXY_SETTING_NAME) == 0)
return _("Proxy");
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* Should not happen; but let's still try to be somewhat sensible. */
return name;
}
static gboolean
questionnaire_one_optional (NmCli *nmc, NMConnection *connection)
{
OptionInfo *candidate;
/* Optional settings. */
const gchar *setting_name = NULL;
int count = 0;
/* Find first setting with relevant options and count them. */
for (candidate = option_info; candidate->setting_name; candidate++) {
if (!option_relevant (connection, candidate))
continue;
if (!setting_name)
setting_name = candidate->setting_name;
else if (strcmp (setting_name, candidate->setting_name))
break;
count++;
Revert "cli: consider empty value as "*" for "Interface name: " prompt" This reverts commit 3cb0bf6bc06293a224bfda4967457af69be89bfc. It broke 'ifname' being required as a mandatory parameter as decided in https://bugzilla.gnome.org/show_bug.cgi?id=698113
2013-12-18 16:05:37 +01:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (!setting_name)
return FALSE;
/* Now ask for the settings. */
if (want_provide_opt_args (setting_name_to_name (setting_name), count)) {
for (candidate = option_info; candidate->setting_name; candidate++) {
if (!option_relevant (connection, candidate))
continue;
if (strcmp (setting_name, candidate->setting_name))
continue;
ask_option (nmc, connection, candidate);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
/* Make sure we won't ask again. */
disable_options (setting_name, NULL);
return TRUE;
}
static NMCResultCode
do_connection_add (NmCli *nmc, int argc, char **argv)
{
NMConnection *connection = NULL;
NMSettingConnection *s_con;
GError *error = NULL;
AddConnectionInfo *info = NULL;
gboolean save_bool = TRUE;
OptionInfo *candidate;
gboolean seen_dash_dash = FALSE;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
rl_attempted_completion_function = (rl_completion_func_t *) nmcli_con_add_tab_completion;
nmc->return_value = NMC_RESULT_SUCCESS;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
/* Create a new connection object */
libnm-core, libnm, core: make NMConnection an interface The fact that NMRemoteConnection has to be an NMConnection and therefore can't be an NMObject means that it needs to reimplement bits of NMObject functionality (and likewise NMObject needs some special magic to deal with it). Likewise, we will need a daemon-side equivalent of NMObject as part of the gdbus port, and we would want NMSettingsConnection to be able to inherit from this as well. Solve this problem by making NMConnection into an interface, and having NMRemoteConnection and NMSettingsConnection implement it. (We use some hacks to keep the GHashTable of NMSettings objects inside nm-connection.c rather than having to be implemented by the implementations.) Since NMConnection is no longer an instantiable type, this adds NMSimpleConnection to replace the various non-D-Bus-based uses of NMConnection throughout the code. nm_connection_new() becomes nm_simple_connection_new(), nm_connection_new_from_hash() becomes nm_simple_connection_new_from_hash(), and nm_connection_duplicate() becomes nm_simple_connection_new_clone().
2014-08-13 14:34:29 -04:00
connection = nm_simple_connection_new ();
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
/* Build up the 'connection' setting */
s_con = (NMSettingConnection *) nm_setting_connection_new ();
nm_connection_add_setting (connection, NM_SETTING (s_con));
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
read_properties:
nmcli: manage "--ask connection add" with no args When parsing arguments of "connection add" we first read the available property-value pairs and then check if the --ask option was passed in order to aid in the fill process of a new connection. Anyway, if there are no property-value tuples at all, we don't even check the --ask option, returning with error. Fix this just checking if any arg is there (argc) before invoking read_connection_properties().
2016-06-22 12:09:37 +02:00
/* Get the arguments from the command line if any */
cli/connection: export read_connection_properties() It makes sense to modify the applied connection from the device object.
2016-06-22 17:35:16 +02:00
if (argc && !nmc_read_connection_properties (nmc, connection, &argc, &argv, &error)) {
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (g_strcmp0 (*argv, "--") == 0 && !seen_dash_dash) {
/* This is for compatibility with older nmcli that required
* options and properties to be separated with "--" */
g_clear_error (&error);
seen_dash_dash = TRUE;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
goto read_properties;
} else if (g_strcmp0 (*argv, "save") == 0) {
/* It would be better if "save" was a separate argument and not
* mixed with properties, but there's not much we can do about it now. */
g_clear_error (&error);
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
g_string_printf (nmc->return_text,
_("Error: value for '%s' argument is required."),
"save");
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
if (!nmc_string_to_bool (*argv, &save_bool, &error)) {
g_string_printf (nmc->return_text, _("Error: 'save': %s."),
error->message);
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
g_clear_error (&error);
goto finish;
}
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
goto read_properties;
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
g_string_assign (nmc->return_text, error->message);
nmc->return_value = error->code;
g_clear_error (&error);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
goto finish;
}
if (nmc->complete)
goto finish;
/* Now ask user for the rest of the mandatory options. */
if (nmc->ask)
questionnaire_mandatory (nmc, connection);
/* Traditionally, we didn't ask for these options for ethernet slaves. They don't
* make much sense, since these are likely to be set by the master anyway. */
if (nm_setting_connection_get_slave_type (s_con)) {
disable_options (NM_SETTING_WIRED_SETTING_NAME, NM_SETTING_WIRED_MTU);
disable_options (NM_SETTING_WIRED_SETTING_NAME, NM_SETTING_WIRED_MAC_ADDRESS);
disable_options (NM_SETTING_WIRED_SETTING_NAME, NM_SETTING_WIRED_CLONED_MAC_ADDRESS);
}
/* Connection id is special in that it's required but we don't insist
* on getting it from the user -- we just make up something sensible. */
if (!nm_setting_connection_get_id (s_con)) {
const char *ifname = nm_setting_connection_get_interface_name (s_con);
const char *type = nm_setting_connection_get_connection_type (s_con);
const char *slave_type = nm_setting_connection_get_slave_type (s_con);
char *try_name, *default_name;
/* If only bother when there's a type, which is not guaranteed at this point.
* Otherwise the validation will fail anyway. */
if (type) {
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
connections = nm_client_get_connections (nmc->client);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
try_name = ifname ? g_strdup_printf ("%s-%s", get_name_alias (type, slave_type, nmc_valid_connection_types), ifname)
: g_strdup (get_name_alias (type, slave_type, nmc_valid_connection_types));
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
default_name = nmc_unique_connection_name (connections, try_name);
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
g_free (try_name);
g_object_set (s_con, NM_SETTING_CONNECTION_ID, default_name, NULL);
g_free (default_name);
}
}
/* For some software connection types we generate the interface name for the user. */
set_default_interface_name (nmc, s_con);
/* Now see if there's something optional that needs to be asked for.
* Keep asking until there's no more things to ask for. */
do {
/* This ensures all settings that make sense are present. */
nm_connection_normalize (connection, NULL, NULL, NULL);
} while (nmc->ask && questionnaire_one_optional (nmc, connection));
/* Mandatory settings. No good reason to check this other than guarding the user
* from doing something that's not likely to make sense (such as missing ifname
* on a bond/bridge/team, etc.). Added just to preserve traditional behavior, it
* perhaps is a good idea to just remove this. */
for (candidate = option_info; candidate->setting_name; candidate++) {
if (!option_relevant (connection, candidate))
continue;
if (candidate->flags & OPTION_REQD) {
g_string_printf (nmc->return_text, _("Error: '%s' argument is required."), candidate->option);
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
}
cli: turn should_wait into a semaphore Count the exit blockers. We'll want to terminate the device monitor in case all monitored devices vanish.
2015-03-27 12:39:17 +01:00
nmc->should_wait++;
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
info = g_malloc0 (sizeof (AddConnectionInfo));
info->nmc = nmc;
info->con_name = g_strdup (nm_connection_get_id (connection));
/* Tell the settings service to add the new connection */
cli: allow creating temporary (unsaved) connection with 'nmcli con add'
2014-01-24 11:04:17 +01:00
add_new_connection (save_bool,
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
nmc->client,
cli: allow creating temporary (unsaved) connection with 'nmcli con add'
2014-01-24 11:04:17 +01:00
connection,
add_connection_cb,
info);
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
finish:
reset_options ();
cli: 'connection add' command for adding NM connections non-interactively A few examples: nmcli --ask connection add nmcli connection add type ethernet nmcli -p connection add type ethernet con-name "my ethernet connection 1" nmcli connection add type ethernet ip4 192.168.100.5/24 nmcli connection add type ethernet ip4 192.168.100.100/24 gw4 192.168.100.1 ip4 1.2.3.4 ip6 abbe::cafe nmcli connection add type ethernet ifname eth0 nmcli connection add type ethernet autoconnect no ifname eth0 nmcli connection add type wifi con-name Rakosnicek ssid Brcalnik nmcli --ask connection add type wifi con-name My_WiFi nmcli c a type infiniband con-name Infi1 transport-mode connected nmcli c a type bluetooth addr 12:54:00:fd:db:26 nmcli c a type bluetooth addr 12:54:00:fd:db:26 bt-type dun-cdma nmcli c a type gsm apn internet nmcli c a type cdma nmcli c a type bond nmcli c a type bond ifname mybond0 mode active-backup nmcli c a ifname maxipes-fik type vlan dev eth0 id 55 nmcli c a con-name VLAN1 type vlan dev eth0 id 44 ingress "2:4,3:55" egress "1:3" flags 6 nmcli c a type bridge ifname br0 stp on ip4 10.0.0.25 nmcli c a ifname eth0 type bridge-slave master 30fc816a-e7dd-4ae7-a86e-ab0c9cee51c2 hairpin no path-cost 333 priority 20 nmcli c a type bridge-slave master br2
2013-04-12 15:26:31 +02:00
if (connection)
g_object_unref (connection);
return nmc->return_value;
}
cli: line-editing capability for interactive editor using an edit line library Supported libraries are: GNU Readline (libreadline) http://cnswww.cns.cwru.edu/php/chet/readline/rltop.html NetBSD Editline (libedit) http://www.thrysoee.dk/editline/ We load these symbols: readline() - main function allowing getting user input and aditing that add_history() - stores strings to history rl_insert_text() - insert text into the line at the current cursor position rl_startup_hook - hook to call just before readline prints the first prompt
2013-01-31 19:03:51 +01:00
/*----------------------------------------------------------------------------*/
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
/* Functions for readline TAB completion in editor */
cli: insert text to line so that user can edit it 'change' command puts current value on edit line, so that user can change it and confirm with <Enter>. Notes: * it works for libreadline * libedit has to be checked * doesn't work without any edit-line library - how to do that??
2013-03-20 15:42:07 +01:00
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
static void
uuid_display_hook (char **array, int len, int max_len)
{
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
NMConnection *con;
int i, max = 0;
char *tmp;
const char *id;
for (i = 1; i <= len; i++) {
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc_tab_completion.nmc->client);
con = nmc_find_connection (connections, "uuid", array[i], NULL, FALSE);
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
id = con ? nm_connection_get_id (con) : NULL;
if (id) {
tmp = g_strdup_printf ("%s (%s)", array[i], id);
g_free (array[i]);
array[i] = tmp;
if (max < strlen (id))
max = strlen (id);
}
}
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_display_match_list (array, len, max_len + max + 3);
rl_forced_update_display ();
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
}
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_nmcli_cmds_menu (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
const char *words[] = { "goto", "set", "remove", "describe", "print", "verify",
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
"save", "activate", "back", "help", "quit", "nmcli",
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_nmcli_cmds_submenu (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
const char *words[] = { "set", "add", "change", "remove", "describe",
"print", "back", "help", "quit",
NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_cmd_nmcli (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
cli: allow showing secrets in the editor
2014-09-29 13:57:22 +02:00
const char *words[] = { "status-line", "save-confirmation", "show-secrets", "prompt-color", NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
}
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_cmd_nmcli_prompt_color (const char *text, int state)
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
{
cli: use color names when TAB-completing "nmcli prompt-color"
2015-06-10 17:47:46 +02:00
const char *words[] = { "normal", "black", "red", "green", "yellow",
"blue", "magenta", "cyan", "white", NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
}
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_func_bool_values (const char *text, int state)
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
{
const char *words[] = { "yes", "no", NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_cmd_verify0 (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
const char *words[] = { "all", "fix", NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
cli: editor - enhance TAB completion for 'print' command
2014-06-18 15:30:29 +02:00
static char *
gen_cmd_print0 (const char *text, int state)
{
static char **words = NULL;
char *ret = NULL;
if (!state) {
libnm-core: change connection hash tables to variants in API In preparation for porting to GDBus, make nm_connection_to_dbus(), etc, represent connections as GVariants of type 'a{sa{sv}}' rather than as GHashTables-of-GHashTables-of-GValues. This means we're constantly converting back and forth internally, but this is just a stepping stone on the way to the full GDBus port, and all of that code will go away again later.
2014-08-16 10:09:48 -04:00
GVariant *settings;
GVariantIter iter;
cli: editor - enhance TAB completion for 'print' command
2014-06-18 15:30:29 +02:00
const char *setting_name;
int i = 0;
libnm-core: rename NMConnection to/from_hash methods Rename nm_connection_to_hash() to nm_connection_to_dbus(), and nm_connection_new_from_hash() to nm_connection_new_from_dbus(). In addition to clarifying that this is specifically the D-Bus serialization format, these names will also work better in the GDBus-based future where the serialization format is GVariant, not GHashTable. Also, move NMSettingHashFlags to nm-connection.h, and rename it NMConnectionSerializationFlags.
2014-08-06 19:35:31 -04:00
settings = nm_connection_to_dbus (nmc_tab_completion.connection, NM_CONNECTION_SERIALIZE_NO_SECRETS);
libnm-core: change connection hash tables to variants in API In preparation for porting to GDBus, make nm_connection_to_dbus(), etc, represent connections as GVariants of type 'a{sa{sv}}' rather than as GHashTables-of-GHashTables-of-GValues. This means we're constantly converting back and forth internally, but this is just a stepping stone on the way to the full GDBus port, and all of that code will go away again later.
2014-08-16 10:09:48 -04:00
words = g_new (char *, g_variant_n_children (settings) + 2);
g_variant_iter_init (&iter, settings);
while (g_variant_iter_next (&iter, "{&s@a{sv}}", &setting_name, NULL))
cli: editor - enhance TAB completion for 'print' command
2014-06-18 15:30:29 +02:00
words [i++] = g_strdup (setting_name);
words[i++] = g_strdup ("all");
words[i] = NULL;
libnm-core: change connection hash tables to variants in API In preparation for porting to GDBus, make nm_connection_to_dbus(), etc, represent connections as GVariants of type 'a{sa{sv}}' rather than as GHashTables-of-GHashTables-of-GValues. This means we're constantly converting back and forth internally, but this is just a stepping stone on the way to the full GDBus port, and all of that code will go away again later.
2014-08-16 10:09:48 -04:00
g_variant_unref (settings);
cli: editor - enhance TAB completion for 'print' command
2014-06-18 15:30:29 +02:00
}
if (words) {
ret = nmc_rl_gen_func_basic (text, state, (const char **) words);
if (ret == NULL) {
g_strfreev (words);
words = NULL;
}
}
return ret;
}
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_cmd_print2 (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
const char *words[] = { "setting", "connection", "all", NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_cmd_save (const char *text, int state)
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
{
const char *words[] = { "persistent", "temporary", NULL };
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
return nmc_rl_gen_func_basic (text, state, words);
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
}
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_connection_types (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
static int list_idx, len;
cli: fix TAB-completion for aliases in interactive editor Before the change if an alias for a setting name existed, it was used instead of the real name and TAB-completion for the real name didn't work. before: nmcli> goto 802<TAB> no output now: nmcli> goto 802<TAB> nmcli> goto 802-11-wireless 802-11-wireless 802-11-wireless-security nmcli> goto <TAB> 802-11-wireless-security (wifi-sec) connection ipv6 802-11-wireless (wifi) ipv4
2013-08-29 15:29:34 +02:00
const char *c_type, *a_type;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
if (!state) {
list_idx = 0;
len = strlen (text);
}
cli: fix TAB-completion for aliases in interactive editor Before the change if an alias for a setting name existed, it was used instead of the real name and TAB-completion for the real name didn't work. before: nmcli> goto 802<TAB> no output now: nmcli> goto 802<TAB> nmcli> goto 802-11-wireless 802-11-wireless 802-11-wireless-security nmcli> goto <TAB> 802-11-wireless-security (wifi-sec) connection ipv6 802-11-wireless (wifi) ipv4
2013-08-29 15:29:34 +02:00
while (nmc_valid_connection_types[list_idx].name) {
a_type = nmc_valid_connection_types[list_idx].alias;
c_type = nmc_valid_connection_types[list_idx].name;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
list_idx++;
cli: fix TAB-completion for aliases in interactive editor Before the change if an alias for a setting name existed, it was used instead of the real name and TAB-completion for the real name didn't work. before: nmcli> goto 802<TAB> no output now: nmcli> goto 802<TAB> nmcli> goto 802-11-wireless 802-11-wireless 802-11-wireless-security nmcli> goto <TAB> 802-11-wireless-security (wifi-sec) connection ipv6 802-11-wireless (wifi) ipv4
2013-08-29 15:29:34 +02:00
if (a_type && !strncmp (text, a_type, len))
return g_strdup (a_type);
if (c_type && !strncmp (text, c_type, len))
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
return g_strdup (c_type);
}
cli: fix TAB-completion for aliases in interactive editor Before the change if an alias for a setting name existed, it was used instead of the real name and TAB-completion for the real name didn't work. before: nmcli> goto 802<TAB> no output now: nmcli> goto 802<TAB> nmcli> goto 802-11-wireless 802-11-wireless 802-11-wireless-security nmcli> goto <TAB> 802-11-wireless-security (wifi-sec) connection ipv6 802-11-wireless (wifi) ipv4
2013-08-29 15:29:34 +02:00
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
return NULL;
}
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_setting_names (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
static int list_idx, len, is_slv;
cli: fix TAB-completion for aliases in interactive editor Before the change if an alias for a setting name existed, it was used instead of the real name and TAB-completion for the real name didn't work. before: nmcli> goto 802<TAB> no output now: nmcli> goto 802<TAB> nmcli> goto 802-11-wireless 802-11-wireless 802-11-wireless-security nmcli> goto <TAB> 802-11-wireless-security (wifi-sec) connection ipv6 802-11-wireless (wifi) ipv4
2013-08-29 15:29:34 +02:00
const char *s_name, *a_name;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
const NameItem *valid_settings_arr;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
NMSettingConnection *s_con;
const char *s_type = NULL;
char *slv_type;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
if (!state) {
list_idx = 0;
len = strlen (text);
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
is_slv = 0;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
if (!is_slv) {
valid_settings_arr = get_valid_settings_array (nmc_tab_completion.con_type);
if (!valid_settings_arr)
return NULL;
while (valid_settings_arr[list_idx].name) {
a_name = valid_settings_arr[list_idx].alias;
s_name = valid_settings_arr[list_idx].name;
list_idx++;
if (len == 0 && a_name)
return g_strdup_printf ("%s (%s)", s_name, a_name);
if (a_name && !strncmp (text, a_name, len))
return g_strdup (a_name);
if (s_name && !strncmp (text, s_name, len))
return g_strdup (s_name);
}
/* Let's give a try to parameters related to slave type */
list_idx = 0;
is_slv = 1;
}
/* is_slv */
s_con = nm_connection_get_setting_connection (nmc_tab_completion.connection);
if (s_con)
s_type = nm_setting_connection_get_slave_type (s_con);
slv_type = g_strdup_printf ("%s-slave", s_type ? s_type : "no");
valid_settings_arr = get_valid_settings_array (slv_type);
g_free (slv_type);
cli: fix TAB-completion for aliases in interactive editor Before the change if an alias for a setting name existed, it was used instead of the real name and TAB-completion for the real name didn't work. before: nmcli> goto 802<TAB> no output now: nmcli> goto 802<TAB> nmcli> goto 802-11-wireless 802-11-wireless 802-11-wireless-security nmcli> goto <TAB> 802-11-wireless-security (wifi-sec) connection ipv6 802-11-wireless (wifi) ipv4
2013-08-29 15:29:34 +02:00
while (valid_settings_arr[list_idx].name) {
a_name = valid_settings_arr[list_idx].alias;
s_name = valid_settings_arr[list_idx].name;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
list_idx++;
cli: fix TAB-completion for aliases in interactive editor Before the change if an alias for a setting name existed, it was used instead of the real name and TAB-completion for the real name didn't work. before: nmcli> goto 802<TAB> no output now: nmcli> goto 802<TAB> nmcli> goto 802-11-wireless 802-11-wireless 802-11-wireless-security nmcli> goto <TAB> 802-11-wireless-security (wifi-sec) connection ipv6 802-11-wireless (wifi) ipv4
2013-08-29 15:29:34 +02:00
if (len == 0 && a_name)
return g_strdup_printf ("%s (%s)", s_name, a_name);
if (a_name && !strncmp (text, a_name, len))
return g_strdup (a_name);
if (s_name && !strncmp (text, s_name, len))
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
return g_strdup (s_name);
}
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
return NULL;
}
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_property_names (const char *text, int state)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
NMSetting *setting = NULL;
char **valid_props = NULL;
char *ret = NULL;
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
const char *line = rl_line_buffer;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
const char *setting_name;
char **strv = NULL;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const NameItem *valid_settings_main;
const NameItem *valid_settings_slave;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
const char *p1;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const char *slv_type;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
/* Try to get the setting from 'line' - setting_name.property */
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
p1 = strchr (line, '.');
if (p1) {
while (p1 > line && !g_ascii_isspace (*p1))
p1--;
strv = g_strsplit (p1+1, ".", 2);
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
valid_settings_main = get_valid_settings_array (nmc_tab_completion.con_type);
/* Support autocompletion of slave-connection parameters
* guessing the slave type from the setting name already
* typed (or autocompleted) */
if (nm_streq0 (strv[0], NM_SETTING_TEAM_PORT_SETTING_NAME))
slv_type = "team-slave";
else if (nm_streq0 (strv[0], NM_SETTING_BRIDGE_PORT_SETTING_NAME))
slv_type = "bridge-slave";
else
slv_type = "no-slave";
valid_settings_slave = get_valid_settings_array (slv_type);
setting_name = check_valid_name (strv[0],
valid_settings_main,
valid_settings_slave,
NULL);
cli: simplify setting creation
2013-06-05 16:04:06 -05:00
setting = nmc_setting_new_for_name (setting_name);
} else {
/* Else take the current setting, if any */
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
setting = nmc_tab_completion.setting ? g_object_ref (nmc_tab_completion.setting) : NULL;
cli: simplify setting creation
2013-06-05 16:04:06 -05:00
}
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
cli: simplify setting creation
2013-06-05 16:04:06 -05:00
if (setting) {
valid_props = nmc_setting_get_valid_properties (setting);
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
ret = nmc_rl_gen_func_basic (text, state, (const char **) valid_props);
cli: simplify setting creation
2013-06-05 16:04:06 -05:00
}
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
g_strfreev (strv);
g_strfreev (valid_props);
if (setting)
g_object_unref (setting);
return ret;
}
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_compat_devices (const char *text, int state)
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
{
int i, j = 0;
const GPtrArray *devices;
const char **compatible_devices;
char *ret;
devices = nm_client_get_devices (nmc_tab_completion.nmc->client);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
if (devices->len == 0)
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
return NULL;
compatible_devices = g_new (const char *, devices->len + 1);
for (i = 0; i < devices->len; i++) {
NMDevice *dev = g_ptr_array_index (devices, i);
cli: fix completion of interfaces in editor for "activate" command
2014-05-15 13:41:31 +02:00
const char *ifname = nm_device_get_iface (dev);
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
NMDevice *device = NULL;
const char *spec_object = NULL;
if (find_device_for_connection (nmc_tab_completion.nmc, nmc_tab_completion.connection,
ifname, NULL, NULL, &device, &spec_object, NULL)) {
compatible_devices[j++] = ifname;
}
}
compatible_devices[j] = NULL;
cli: move readline related functions to common.h to be usable throughout nmcli Renames readline_x() to nmc_readline() gen_func_basic() to nmc_rl_gen_func_basic() The commit doesn't change functionality, only moves and renames the functions, so that they can be used in other places too.
2014-05-15 10:31:09 +02:00
ret = nmc_rl_gen_func_basic (text, state, compatible_devices);
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
g_free (compatible_devices);
return ret;
}
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
static const char **
_create_vpn_array (const GPtrArray *connections, gboolean uuid)
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
{
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
int c, idx = 0;
const char **array;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
if (connections->len < 1)
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
return NULL;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
array = g_new (const char *, connections->len + 1);
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
for (c = 0; c < connections->len; c++) {
NMConnection *connection = NM_CONNECTION (connections->pdata[c]);
const char *type = nm_connection_get_connection_type (connection);
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
if (g_strcmp0 (type, NM_SETTING_VPN_SETTING_NAME) == 0)
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
array[idx++] = uuid ? nm_connection_get_uuid (connection) : nm_connection_get_id (connection);
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
}
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
array[idx] = NULL;
return array;
}
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
static char *
gen_vpn_uuids (const char *text, int state)
{
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
const char **uuids;
char *ret;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nm_cli.client);
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
if (connections->len < 1)
return NULL;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
uuids = _create_vpn_array (connections, TRUE);
ret = nmc_rl_gen_func_basic (text, state, uuids);
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
g_free (uuids);
return ret;
}
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
static char *
gen_vpn_ids (const char *text, int state)
{
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
const char **ids;
char *ret;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nm_cli.client);
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
if (connections->len < 1)
return NULL;
ids = _create_vpn_array (connections, FALSE);
ret = nmc_rl_gen_func_basic (text, state, ids);
g_free (ids);
return ret;
}
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
static rl_compentry_func_t *
get_gen_func_cmd_nmcli (const char *str)
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
{
if (!str)
return NULL;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (str, "status-line"))
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
return gen_func_bool_values;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (str, "save-confirmation"))
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
return gen_func_bool_values;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (str, "show-secrets"))
cli: allow showing secrets in the editor
2014-09-29 13:57:22 +02:00
return gen_func_bool_values;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (str, "prompt-color"))
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
return gen_cmd_nmcli_prompt_color;
return NULL;
}
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
/*
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
* Helper function parsing line for completion.
* IN:
* line : the whole line to be parsed
* end : the position of cursor in the line
* cmd : command to match
* OUT:
* cw_num : is set to the word number being completed (1, 2, 3, 4).
* prev_word : returns the previous word (so that we have some context).
*
* Returns TRUE when the first word of the 'line' matches 'cmd'.
*
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
* Examples:
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
* line="rem" cmd="remove" -> TRUE cw_num=1
* line="set con" cmd="set" -> TRUE cw_num=2
* line="go ipv4.method" cmd="goto" -> TRUE cw_num=2
* line=" des eth.mtu " cmd="describe" -> TRUE cw_num=3
* line=" bla ipv4.method" cmd="goto" -> FALSE
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
*/
static gboolean
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
should_complete_cmd (const char *line, int end, const char *cmd,
int *cw_num, char **prev_word)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
char *tmp;
const char *word1, *word2, *word3;
size_t n1, n2, n3, n4, n5, n6;
gboolean word1_done, word2_done, word3_done;
gboolean ret = FALSE;
if (!line)
return FALSE;
tmp = g_strdup (line);
n1 = strspn (tmp, " \t");
n2 = strcspn (tmp+n1, " \t\0") + n1;
n3 = strspn (tmp+n2, " \t") + n2;
n4 = strcspn (tmp+n3, " \t\0") + n3;
n5 = strspn (tmp+n4, " \t") + n4;
n6 = strcspn (tmp+n5, " \t\0") + n5;
word1_done = end > n2;
word2_done = end > n4;
word3_done = end > n6;
tmp[n2] = tmp[n4] = tmp[n6] = '\0';
word1 = tmp[n1] ? tmp + n1 : NULL;
word2 = tmp[n3] ? tmp + n3 : NULL;
word3 = tmp[n5] ? tmp + n5 : NULL;
if (!word1_done) {
if (cw_num)
*cw_num = 1;
if (prev_word)
*prev_word = NULL;
} else if (!word2_done) {
if (cw_num)
*cw_num = 2;
if (prev_word)
*prev_word = g_strdup (word1);
} else if (!word3_done) {
if (cw_num)
*cw_num = 3;
if (prev_word)
*prev_word = g_strdup (word2);
} else {
if (cw_num)
*cw_num = 4;
if (prev_word)
*prev_word = g_strdup (word3);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (word1 && matches (word1, cmd))
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
ret = TRUE;
g_free (tmp);
return ret;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
}
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
/*
* extract_setting_and_property:
* prompt: (in) (allow-none): prompt string, or NULL
* line: (in) (allow-none): line, or NULL
* setting: (out) (transfer full) (array zero-terminated=1):
* return location for setting name
* property: (out) (transfer full) (array zero-terminated=1):
* return location for property name
*
* Extract setting and property names from prompt and/or line.
*/
static void
extract_setting_and_property (const char *prompt, const char *line,
char **setting, char **property)
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
{
char *prop = NULL;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
char *sett = NULL;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
if (prompt) {
/* prompt looks like this:
"nmcli 802-1x>" or "nmcli 802-1x.pac-file>" */
const char *p1, *p2, *dot;
size_t num1, num2;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
p1 = strchr (prompt, ' ');
if (p1) {
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
dot = strchr (++p1, '.');
if (dot) {
p2 = dot + 1;
num1 = strcspn (p1, ".");
num2 = strcspn (p2, ">");
sett = num1 > 0 ? g_strndup (p1, num1) : NULL;
prop = num2 > 0 ? g_strndup (p2, num2) : NULL;
} else {
num1 = strcspn (p1, ">");
sett = num1 > 0 ? g_strndup (p1, num1) : NULL;
}
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
}
}
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
if (line) {
/* line looks like this:
" set 802-1x.pac-file ..." or " set pac-file ..." */
const char *p1, *p2, *dot;
size_t n1, n2, n3, n4;
size_t num1, num2, len;
n1 = strspn (line, " \t"); /* white-space */
n2 = strcspn (line+n1, " \t\0") + n1; /* command */
n3 = strspn (line+n2, " \t") + n2; /* white-space */
n4 = strcspn (line+n3, " \t\0") + n3; /* setting/property */
p1 = line + n3;
len = n4 - n3;
dot = strchr (p1, '.');
if (dot && dot < p1 + len) {
p2 = dot + 1;
num1 = strcspn (p1, ".");
num2 = len > num1 + 1 ? len - num1 - 1 : 0;
sett = num1 > 0 ? g_strndup (p1, num1) : sett;
prop = num2 > 0 ? g_strndup (p2, num2) : prop;
} else {
if (!prop)
prop = len > 0 ? g_strndup (p1, len) : NULL;
}
}
if (setting)
*setting = sett;
else
g_free (sett);
if (property)
*property = prop;
else
g_free (prop);
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
}
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
static void
get_setting_and_property (const char *prompt, const char *line,
NMSetting **setting_out, char**property_out)
{
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const NameItem *valid_settings_main;
const NameItem *valid_settings_slave;
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
const char *setting_name;
NMSetting *setting = NULL;
char *property = NULL;
char *sett = NULL, *prop = NULL;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
NMSettingConnection *s_con;
const char *s_type = NULL;
char *slv_type;
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
extract_setting_and_property (prompt, line, &sett, &prop);
if (sett) {
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
/* Is this too much (and useless?) effort for an unlikely case? */
s_con = nm_connection_get_setting_connection (nmc_tab_completion.connection);
if (s_con)
s_type = nm_setting_connection_get_slave_type (s_con);
slv_type = g_strdup_printf ("%s-slave", s_type ? s_type : "no");
valid_settings_main = get_valid_settings_array (nmc_tab_completion.con_type);
valid_settings_slave = get_valid_settings_array (slv_type);
g_free (slv_type);
setting_name = check_valid_name (sett, valid_settings_main,
valid_settings_slave, NULL);
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
setting = nmc_setting_new_for_name (setting_name);
} else
setting = nmc_tab_completion.setting ? g_object_ref (nmc_tab_completion.setting) : NULL;
if (setting && prop)
property = is_property_valid (setting, prop, NULL);
else
property = g_strdup (nmc_tab_completion.property);
*setting_out = setting;
*property_out = property;
g_free (sett);
g_free (prop);
}
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
static gboolean
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
_get_and_check_property (const char *prompt,
const char *line,
const char **array,
const char **array_multi,
gboolean *multi)
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
{
char *prop;
gboolean found = FALSE;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
extract_setting_and_property (prompt, line, NULL, &prop);
if (prop) {
if (array)
found = !!nmc_string_is_valid (prop, array, NULL);
if (array_multi && multi)
*multi = !!nmc_string_is_valid (prop, array_multi, NULL);
g_free (prop);
}
return found;
}
static gboolean
should_complete_files (const char *prompt, const char *line)
{
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
const char *file_properties[] = {
/* '802-1x' properties */
"ca-cert",
"ca-path",
"client-cert",
"pac-file",
"phase2-ca-cert",
"phase2-ca-path",
"phase2-client-cert",
"private-key",
"phase2-private-key",
/* 'team' and 'team-port' properties */
"config",
cli: enable TAB-completion for proxy properties in interactive editor
2017-01-17 10:18:00 +01:00
/* 'proxy' properties */
"pac-script",
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
NULL
};
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
return _get_and_check_property (prompt, line, file_properties, NULL, NULL);
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
}
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
static gboolean
should_complete_vpn_uuids (const char *prompt, const char *line)
{
const char *uuid_properties[] = {
/* 'connection' properties */
"secondaries",
NULL
};
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
return _get_and_check_property (prompt, line, uuid_properties, NULL, NULL);
}
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
cli: add values_func() to NmcPropertyInfo and use for "connection" setting
2017-03-25 13:20:11 +01:00
static const char *const*
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
get_allowed_property_values (char ***out_to_free)
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
{
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_unref_object NMSetting *setting = NULL;
gs_free char *property = NULL;
cli: add values_func() to NmcPropertyInfo and use for "connection" setting
2017-03-25 13:20:11 +01:00
const char *const*avals = NULL;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
get_setting_and_property (rl_prompt, rl_line_buffer, &setting, &property);
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
if (setting && property)
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
avals = nmc_setting_get_property_allowed_values (setting, property, out_to_free);
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
return avals;
}
static gboolean
should_complete_property_values (const char *prompt, const char *line, gboolean *multi)
{
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_strfreev char **to_free = NULL;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
/* properties allowing multiple values */
const char *multi_props[] = {
/* '802-1x' properties */
NM_SETTING_802_1X_EAP,
/* '802-11-wireless-security' properties */
NM_SETTING_WIRELESS_SECURITY_PROTO,
NM_SETTING_WIRELESS_SECURITY_PAIRWISE,
NM_SETTING_WIRELESS_SECURITY_GROUP,
/* 'bond' properties */
NM_SETTING_BOND_OPTIONS,
/* 'ethernet' properties */
NM_SETTING_WIRED_S390_OPTIONS,
NULL
};
_get_and_check_property (prompt, line, NULL, multi_props, multi);
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
return !!get_allowed_property_values (&to_free);
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
}
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
static gboolean
_setting_property_is_boolean (NMSetting *setting, const char *property_name)
{
GParamSpec *pspec;
g_return_val_if_fail (NM_IS_SETTING (setting), FALSE);
g_return_val_if_fail (property_name, FALSE);
pspec = g_object_class_find_property (G_OBJECT_GET_CLASS (setting), property_name);
if (pspec && pspec->value_type == G_TYPE_BOOLEAN)
return TRUE;
return FALSE;
}
static gboolean
should_complete_boolean (const char *prompt, const char *line)
{
NMSetting *setting;
char *property;
gboolean is_boolean = FALSE;
get_setting_and_property (prompt, line, &setting, &property);
if (setting && property)
is_boolean = _setting_property_is_boolean (setting, property);
if (setting)
g_object_unref (setting);
g_free (property);
return is_boolean;
}
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
static char *
gen_property_values (const char *text, int state)
{
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_strfreev char **to_free = NULL;
cli: add values_func() to NmcPropertyInfo and use for "connection" setting
2017-03-25 13:20:11 +01:00
const char *const*avals;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
avals = get_allowed_property_values (&to_free);
if (!avals)
return NULL;
return nmc_rl_gen_func_basic (text, state, avals);
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
}
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
/* from readline */
extern int rl_complete_with_tilde_expansion;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
/*
* Attempt to complete on the contents of TEXT. START and END show the
* region of TEXT that contains the word to complete. We can use the
* entire line in case we want to do some simple parsing. Return the
* array of matches, or NULL if there aren't any.
*/
static char **
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
nmcli_editor_tab_completion (const char *text, int start, int end)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
{
char **match_array = NULL;
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
const char *line = rl_line_buffer;
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
rl_compentry_func_t *generator_func = NULL;
nmcli: make filtering color sequences a utility function The code will be used on other places too.
2015-02-13 13:25:16 +01:00
char *prompt_tmp;
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
char *word = NULL;
size_t n1;
int num;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
/* Restore standard append character to space */
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_completion_append_character = ' ';
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
/* Restore standard function for displaying matches */
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_completion_display_matches_hook = NULL;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
/* Disable default filename completion */
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_attempted_completion_over = 1;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
/* Enable tilde expansion when filenames are completed */
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_complete_with_tilde_expansion = 1;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
/* Filter out possible ANSI color escape sequences */
nmcli: make filtering color sequences a utility function The code will be used on other places too.
2015-02-13 13:25:16 +01:00
prompt_tmp = nmc_filter_out_colors ((const char *) rl_prompt);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
/* Find the first non-space character */
n1 = strspn (line, " \t");
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
/* Choose the right generator function */
if (strcmp (prompt_tmp, EDITOR_PROMPT_CON_TYPE) == 0)
generator_func = gen_connection_types;
else if (strcmp (prompt_tmp, EDITOR_PROMPT_SETTING) == 0)
generator_func = gen_setting_names;
else if (strcmp (prompt_tmp, EDITOR_PROMPT_PROPERTY) == 0)
generator_func = gen_property_names;
cli: use readline for quit confirmation in editor, and accept localized answer
2014-06-02 12:24:56 +02:00
else if ( g_str_has_suffix (rl_prompt, prompt_yes_no (TRUE, NULL))
|| g_str_has_suffix (rl_prompt, prompt_yes_no (FALSE, NULL)))
generator_func = gen_func_bool_values_l10n;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
else if (g_str_has_prefix (prompt_tmp, "nmcli")) {
if (!strchr (prompt_tmp, '.')) {
int level = g_str_has_prefix (prompt_tmp, "nmcli>") ? 0 : 1;
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
const char *dot = strchr (line, '.');
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
gboolean multi;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
/* Main menu - level 0,1 */
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
if (start == n1)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_nmcli_cmds_menu;
else {
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
if (should_complete_cmd (line, end, "goto", &num, NULL) && num <= 2) {
if (level == 0 && (!dot || dot >= line + end))
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_setting_names;
else
generator_func = gen_property_names;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
} else if (should_complete_cmd (line, end, "set", &num, NULL)) {
if (num < 3) {
if (level == 0 && (!dot || dot >= line + end)) {
generator_func = gen_setting_names;
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_completion_append_character = '.';
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
} else
generator_func = gen_property_names;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
} else if (num >= 3) {
if (num == 3 && should_complete_files (NULL, line))
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_attempted_completion_over = 0;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
else if (should_complete_vpn_uuids (NULL, line)) {
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_completion_display_matches_hook = uuid_display_hook;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
generator_func = gen_vpn_uuids;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
} else if ( should_complete_property_values (NULL, line, &multi)
&& (num == 3 || multi)) {
generator_func = gen_property_values;
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
} else if (should_complete_boolean (NULL, line) && num == 3)
generator_func = gen_func_bool_values;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
}
} else if ( ( should_complete_cmd (line, end, "remove", &num, NULL)
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
|| should_complete_cmd (line, end, "describe", &num, NULL))
&& num <= 2) {
if (level == 0 && (!dot || dot >= line + end)) {
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_setting_names;
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_completion_append_character = '.';
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
} else
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_property_names;
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
} else if (should_complete_cmd (line, end, "nmcli", &num, &word)) {
if (num < 3)
generator_func = gen_cmd_nmcli;
else if (num == 3)
generator_func = get_gen_func_cmd_nmcli (word);
cli: editor - enhance TAB completion for 'print' command
2014-06-18 15:30:29 +02:00
} else if (should_complete_cmd (line, end, "print", &num, NULL) && num <= 2) {
if (level == 0 && (!dot || dot >= line + end))
generator_func = gen_cmd_print0;
else
generator_func = gen_property_names;
} else if (should_complete_cmd (line, end, "verify", &num, NULL) && num <= 2) {
generator_func = gen_cmd_verify0;
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
} else if (should_complete_cmd (line, end, "activate", &num, NULL) && num <= 2) {
generator_func = gen_compat_devices;
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
} else if (should_complete_cmd (line, end, "save", &num, NULL) && num <= 2) {
generator_func = gen_cmd_save;
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
} else if (should_complete_cmd (line, end, "help", &num, NULL) && num <= 2)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_nmcli_cmds_menu;
}
} else {
/* Submenu - level 2 */
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
if (start == n1)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_nmcli_cmds_submenu;
else {
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
gboolean multi;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
if ( should_complete_cmd (line, end, "add", &num, NULL)
|| should_complete_cmd (line, end, "set", &num, NULL)) {
if (num <= 2 && should_complete_files (prompt_tmp, line))
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_attempted_completion_over = 0;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
else if (should_complete_vpn_uuids (prompt_tmp, line)) {
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
rl_completion_display_matches_hook = uuid_display_hook;
cli: TAB completion for connection.secondaries We complete UUIDs, because that's easier (no spaces or special characters). Readline display hook allows to display profile name too, though.
2014-04-22 16:28:21 +02:00
generator_func = gen_vpn_uuids;
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
} else if ( should_complete_property_values (prompt_tmp, NULL, &multi)
&& (num <= 2 || multi)) {
generator_func = gen_property_values;
cli: TAB-complete yes/no for boolean properties in the editor
2015-05-12 10:33:48 +02:00
} else if (should_complete_boolean (prompt_tmp, NULL) && num <= 2)
generator_func = gen_func_bool_values;
cli: complete file names for properties whose value is a file name Note: Tilde expansion is enabled, so ~/abc<TAB> expands the string to /home/joe/abc. However, when user did not use the completion and typed "~/myfile" manually, the path could not be opened. nmcli 802-1x.ca-cert> set ~/newca.crt Error: failed to set 'ca-cert' property: Failed to open file '~/newca.crt': No such file or directory
2013-10-03 20:20:58 +02:00
}
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
if (should_complete_cmd (line, end, "print", &num, NULL) && num <= 2)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_cmd_print2;
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
else if (should_complete_cmd (line, end, "help", &num, NULL) && num <= 2)
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
generator_func = gen_nmcli_cmds_submenu;
}
}
}
if (generator_func)
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
match_array = rl_completion_matches (text, generator_func);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
g_free (prompt_tmp);
cli: enhance TAB-completion in nmcli editor - completion now works for command shortcuts too nmcli> g ethe<TAB> - completion now allows leading spaces nmcli> desc wifi.ss<TAB> - fix completion of settings when property is already there nmcli> set con<TAB>.id - fix completion of properties in setting.property nmcli> des con.inter<TAB> bla
2013-10-03 09:13:23 +01:00
g_free (word);
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
return match_array;
}
cli: load/save nmcli editor (readline) history so that it is persistent The history is saved to ~/.nmcli-history file, separately for each connection. The file uses glib key-file format. Each group is a connection UUID, keys are integer numbers (history entry order), and values are the actual commands. Example ~/.nmcli-history file: [0bdc9852-2540-4e12-a605-5e65a9483772] 0=help quit 1=print 2=nmcli prompt-color 3 3=help set 4=q [9142680d-3b87-4feb-ab1e-19e8762329ad] 0=eth 1=set ipv4.addr 1.2.3.4 2=quit
2013-06-24 10:29:49 +02:00
#define NMCLI_EDITOR_HISTORY ".nmcli-history"
static void
load_history_cmds (const char *uuid)
{
GKeyFile *kf;
char *filename;
char **keys;
char *line;
size_t i;
GError *err = NULL;
filename = g_build_filename (g_get_home_dir (), NMCLI_EDITOR_HISTORY, NULL);
kf = g_key_file_new ();
if (!g_key_file_load_from_file (kf, filename, G_KEY_FILE_KEEP_COMMENTS, &err)) {
all: use g_error_matches() Use g_error_matches() where we're testing error codes. In particular, use it rather than looking at only ->code and not also ->domain, which is just wrong. [thaller@redhat.com: rebase and modify original patch]
2014-04-08 12:00:09 -04:00
if (g_error_matches (err, G_KEY_FILE_ERROR, G_KEY_FILE_ERROR_PARSE))
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("Warning: %s parse error: %s\n", filename, err->message);
cli: load/save nmcli editor (readline) history so that it is persistent The history is saved to ~/.nmcli-history file, separately for each connection. The file uses glib key-file format. Each group is a connection UUID, keys are integer numbers (history entry order), and values are the actual commands. Example ~/.nmcli-history file: [0bdc9852-2540-4e12-a605-5e65a9483772] 0=help quit 1=print 2=nmcli prompt-color 3 3=help set 4=q [9142680d-3b87-4feb-ab1e-19e8762329ad] 0=eth 1=set ipv4.addr 1.2.3.4 2=quit
2013-06-24 10:29:49 +02:00
g_key_file_free (kf);
g_free (filename);
return;
}
keys = g_key_file_get_keys (kf, uuid, NULL, NULL);
for (i = 0; keys && keys[i]; i++) {
line = g_key_file_get_string (kf, uuid, keys[i], NULL);
if (line && *line)
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
add_history (line);
cli: load/save nmcli editor (readline) history so that it is persistent The history is saved to ~/.nmcli-history file, separately for each connection. The file uses glib key-file format. Each group is a connection UUID, keys are integer numbers (history entry order), and values are the actual commands. Example ~/.nmcli-history file: [0bdc9852-2540-4e12-a605-5e65a9483772] 0=help quit 1=print 2=nmcli prompt-color 3 3=help set 4=q [9142680d-3b87-4feb-ab1e-19e8762329ad] 0=eth 1=set ipv4.addr 1.2.3.4 2=quit
2013-06-24 10:29:49 +02:00
g_free (line);
}
g_strfreev (keys);
g_key_file_free (kf);
g_free (filename);
}
static void
save_history_cmds (const char *uuid)
{
HIST_ENTRY **hist = NULL;
GKeyFile *kf;
char *filename;
size_t i;
char *key;
char *data;
gsize len = 0;
GError *err = NULL;
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
hist = history_list ();
cli: load/save nmcli editor (readline) history so that it is persistent The history is saved to ~/.nmcli-history file, separately for each connection. The file uses glib key-file format. Each group is a connection UUID, keys are integer numbers (history entry order), and values are the actual commands. Example ~/.nmcli-history file: [0bdc9852-2540-4e12-a605-5e65a9483772] 0=help quit 1=print 2=nmcli prompt-color 3 3=help set 4=q [9142680d-3b87-4feb-ab1e-19e8762329ad] 0=eth 1=set ipv4.addr 1.2.3.4 2=quit
2013-06-24 10:29:49 +02:00
if (hist) {
filename = g_build_filename (g_get_home_dir (), NMCLI_EDITOR_HISTORY, NULL);
kf = g_key_file_new ();
if (!g_key_file_load_from_file (kf, filename, G_KEY_FILE_KEEP_COMMENTS, &err)) {
all: use g_error_matches() Use g_error_matches() where we're testing error codes. In particular, use it rather than looking at only ->code and not also ->domain, which is just wrong. [thaller@redhat.com: rebase and modify original patch]
2014-04-08 12:00:09 -04:00
if ( !g_error_matches (err, G_FILE_ERROR, G_FILE_ERROR_NOENT)
&& !g_error_matches (err, G_KEY_FILE_ERROR, G_KEY_FILE_ERROR_NOT_FOUND)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("Warning: %s parse error: %s\n", filename, err->message);
cli: load/save nmcli editor (readline) history so that it is persistent The history is saved to ~/.nmcli-history file, separately for each connection. The file uses glib key-file format. Each group is a connection UUID, keys are integer numbers (history entry order), and values are the actual commands. Example ~/.nmcli-history file: [0bdc9852-2540-4e12-a605-5e65a9483772] 0=help quit 1=print 2=nmcli prompt-color 3 3=help set 4=q [9142680d-3b87-4feb-ab1e-19e8762329ad] 0=eth 1=set ipv4.addr 1.2.3.4 2=quit
2013-06-24 10:29:49 +02:00
g_key_file_free (kf);
g_free (filename);
g_clear_error (&err);
return;
}
g_clear_error (&err);
}
/* Remove previous history group and save new history entries */
g_key_file_remove_group (kf, uuid, NULL);
for (i = 0; hist[i]; i++)
{
key = g_strdup_printf ("%zd", i);
g_key_file_set_string (kf, uuid, key, hist[i]->line);
g_free (key);
}
/* Write history to file */
data = g_key_file_to_data (kf, &len, NULL);
if (data) {
g_file_set_contents (filename, data, len, NULL);
g_free (data);
}
g_key_file_free (kf);
g_free (filename);
}
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/*----------------------------------------------------------------------------*/
static void
editor_show_connection (NMConnection *connection, NmCli *nmc)
{
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
nmc->nmc_config_mutable.print_output = NMC_PRINT_PRETTY;
nmc->nmc_config_mutable.multiline_output = TRUE;
nmc->nmc_config_mutable.escape_values = 0;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: allow showing secrets in the editor
2014-09-29 13:57:22 +02:00
nmc_connection_profile_details (connection, nmc, nmc->editor_show_secrets);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
static void
editor_show_setting (NMSetting *setting, NmCli *nmc)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("['%s' setting values]\n"),
nm_setting_get_name (setting));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
nmc->nmc_config_mutable.print_output = NMC_PRINT_NORMAL;
nmc->nmc_config_mutable.multiline_output = TRUE;
nmc->nmc_config_mutable.escape_values = 0;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: add nmc_output_selection_create() to parse field selection nmc_output_selection_create() returns a less opaque result then a GArray and a GPtrArray for the groups.
2017-04-03 14:04:28 +02:00
setting_details (&nmc->nmc_config, setting, NULL, nmc->editor_show_secrets);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
typedef enum {
NMC_EDITOR_MAIN_CMD_UNKNOWN = 0,
NMC_EDITOR_MAIN_CMD_GOTO,
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
NMC_EDITOR_MAIN_CMD_REMOVE,
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NMC_EDITOR_MAIN_CMD_SET,
NMC_EDITOR_MAIN_CMD_DESCRIBE,
NMC_EDITOR_MAIN_CMD_PRINT,
NMC_EDITOR_MAIN_CMD_VERIFY,
NMC_EDITOR_MAIN_CMD_SAVE,
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
NMC_EDITOR_MAIN_CMD_ACTIVATE,
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NMC_EDITOR_MAIN_CMD_BACK,
NMC_EDITOR_MAIN_CMD_HELP,
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
NMC_EDITOR_MAIN_CMD_NMCLI,
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NMC_EDITOR_MAIN_CMD_QUIT,
} NmcEditorMainCmd;
static NmcEditorMainCmd
parse_editor_main_cmd (const char *cmd, char **cmd_arg)
{
NmcEditorMainCmd editor_cmd = NMC_EDITOR_MAIN_CMD_UNKNOWN;
char **vec;
cli: fix an 'nmcli con edit' crash nmcli> desc AAA
2013-09-11 18:05:32 +02:00
vec = nmc_strsplit_set (cmd, " \t", 2);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (g_strv_length (vec) < 1) {
if (cmd_arg)
*cmd_arg = NULL;
return NMC_EDITOR_MAIN_CMD_UNKNOWN;
}
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (vec[0], "goto"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_GOTO;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "remove"))
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
editor_cmd = NMC_EDITOR_MAIN_CMD_REMOVE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "set"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_SET;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "describe"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_DESCRIBE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "print"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_PRINT;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "verify"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_VERIFY;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "save"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_SAVE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "activate"))
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
editor_cmd = NMC_EDITOR_MAIN_CMD_ACTIVATE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "back"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_BACK;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "help") || strcmp (vec[0], "?") == 0)
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_HELP;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "quit"))
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
editor_cmd = NMC_EDITOR_MAIN_CMD_QUIT;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "nmcli"))
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
editor_cmd = NMC_EDITOR_MAIN_CMD_NMCLI;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* set pointer to command argument */
if (cmd_arg)
cli: don't call g_strstrip() on NULL arguments GLib-CRITICAL **: g_strchug: assertion `string != NULL' failed GLib-CRITICAL **: g_strchomp: assertion `string != NULL' failed
2013-09-12 10:03:23 +02:00
*cmd_arg = vec[1] ? g_strstrip (g_strdup (vec[1])) : NULL;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_strfreev (vec);
return editor_cmd;
}
static void
editor_main_usage (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("------------------------------------------------------------------------------\n");
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* TRANSLATORS: do not translate command names and keywords before ::
* However, you should translate terms enclosed in <>.
*/
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("---[ Main menu ]---\n"
"goto [<setting> | <prop>] :: go to a setting or property\n"
"remove <setting>[.<prop>] | <prop> :: remove setting or reset property value\n"
"set [<setting>.<prop> <value>] :: set property value\n"
"describe [<setting>.<prop>] :: describe property\n"
"print [all | <setting>[.<prop>]] :: print the connection\n"
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
"verify [all | fix] :: verify the connection\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"save [persistent|temporary] :: save the connection\n"
"activate [<ifname>] [/<ap>|<nsp>] :: activate the connection\n"
"back :: go one level up (back)\n"
"help/? [<command>] :: print this help\n"
"nmcli <conf-option> <value> :: nmcli configuration\n"
"quit :: exit nmcli\n"));
g_print ("------------------------------------------------------------------------------\n");
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
static void
editor_main_help (const char *command)
{
if (!command)
editor_main_usage ();
else {
/* detailed command descriptions */
NmcEditorMainCmd cmd = parse_editor_main_cmd (command, NULL);
switch (cmd) {
case NMC_EDITOR_MAIN_CMD_GOTO:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("goto <setting>[.<prop>] | <prop> :: enter setting/property for editing\n\n"
"This command enters into a setting or property for editing it.\n\n"
"Examples: nmcli> goto connection\n"
" nmcli connection> goto secondaries\n"
" nmcli> goto ipv4.addresses\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
case NMC_EDITOR_MAIN_CMD_REMOVE:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("remove <setting>[.<prop>] :: remove setting or reset property value\n\n"
"This command removes an entire setting from the connection, or if a property\n"
"is given, resets that property to the default value.\n\n"
"Examples: nmcli> remove wifi-sec\n"
" nmcli> remove eth.mtu\n"));
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
break;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
case NMC_EDITOR_MAIN_CMD_SET:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("set [<setting>.<prop> <value>] :: set property value\n\n"
"This command sets property value.\n\n"
"Example: nmcli> set con.id My connection\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
case NMC_EDITOR_MAIN_CMD_DESCRIBE:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("describe [<setting>.<prop>] :: describe property\n\n"
"Shows property description. You can consult nm-settings(5) "
"manual page to see all NM settings and properties.\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
case NMC_EDITOR_MAIN_CMD_PRINT:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("print [all] :: print setting or connection values\n\n"
"Shows current property or the whole connection.\n\n"
"Example: nmcli ipv4> print all\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
case NMC_EDITOR_MAIN_CMD_VERIFY:
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
g_print (_("verify [all | fix] :: verify setting or connection validity\n\n"
"Verifies whether the setting or connection is valid and can be saved later.\n"
"It indicates invalid values on error. Some errors may be fixed automatically\n"
"by 'fix' option.\n\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"Examples: nmcli> verify\n"
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
" nmcli> verify fix\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" nmcli bond> verify\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
case NMC_EDITOR_MAIN_CMD_SAVE:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("save [persistent|temporary] :: save the connection\n\n"
"Sends the connection profile to NetworkManager that either will save it\n"
"persistently, or will only keep it in memory. 'save' without an argument\n"
"means 'save persistent'.\n"
"Note that once you save the profile persistently those settings are saved\n"
"across reboot or restart. Subsequent changes can also be temporary or\n"
"persistent, but any temporary changes will not persist across reboot or\n"
"restart. If you want to fully remove the persistent connection, the connection\n"
"profile must be deleted.\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
case NMC_EDITOR_MAIN_CMD_ACTIVATE:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("activate [<ifname>] [/<ap>|<nsp>] :: activate the connection\n\n"
"Activates the connection.\n\n"
"Available options:\n"
"<ifname> - device the connection will be activated on\n"
"/<ap>|<nsp> - AP (Wi-Fi) or NSP (WiMAX) (prepend with / when <ifname> is not specified)\n"));
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
break;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
case NMC_EDITOR_MAIN_CMD_BACK:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("back :: go to upper menu level\n\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
case NMC_EDITOR_MAIN_CMD_HELP:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("help/? [<command>] :: help for the nmcli commands\n\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
case NMC_EDITOR_MAIN_CMD_NMCLI:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("nmcli [<conf-option> <value>] :: nmcli configuration\n\n"
"Configures nmcli. The following options are available:\n"
cli: take color name arguments for "nmcli prompt-color" in editor (bgo #744936) Adjust nmcli prompt-color description and make it more friendly for translators. nmcli> nmcli prompt-color green https://bugzilla.gnome.org/show_bug.cgi?id=744936
2015-06-10 17:41:53 +02:00
"status-line yes | no [default: no]\n"
"save-confirmation yes | no [default: yes]\n"
"show-secrets yes | no [default: no]\n"
"prompt-color <color> | <0-8> [default: 0]\n"
"%s" /* color table description */
"\n"
"Examples: nmcli> nmcli status-line yes\n"
" nmcli> nmcli save-confirmation no\n"
" nmcli> nmcli prompt-color 3\n"),
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
" 0 = normal\n"
" 1 = \33[30mblack\33[0m\n"
" 2 = \33[31mred\33[0m\n"
" 3 = \33[32mgreen\33[0m\n"
" 4 = \33[33myellow\33[0m\n"
" 5 = \33[34mblue\33[0m\n"
" 6 = \33[35mmagenta\33[0m\n"
" 7 = \33[36mcyan\33[0m\n"
cli: take color name arguments for "nmcli prompt-color" in editor (bgo #744936) Adjust nmcli prompt-color description and make it more friendly for translators. nmcli> nmcli prompt-color green https://bugzilla.gnome.org/show_bug.cgi?id=744936
2015-06-10 17:41:53 +02:00
" 8 = \33[37mwhite\33[0m\n");
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
break;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
case NMC_EDITOR_MAIN_CMD_QUIT:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("quit :: exit nmcli\n\n"
"This command exits nmcli. When the connection being edited "
"is not saved, the user is asked to confirm the action.\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
default:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Unknown command: '%s'\n"), command);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
}
}
}
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
typedef enum {
NMC_EDITOR_SUB_CMD_UNKNOWN = 0,
NMC_EDITOR_SUB_CMD_SET,
NMC_EDITOR_SUB_CMD_ADD,
NMC_EDITOR_SUB_CMD_CHANGE,
NMC_EDITOR_SUB_CMD_REMOVE,
NMC_EDITOR_SUB_CMD_DESCRIBE,
NMC_EDITOR_SUB_CMD_PRINT,
NMC_EDITOR_SUB_CMD_BACK,
NMC_EDITOR_SUB_CMD_HELP,
NMC_EDITOR_SUB_CMD_QUIT
} NmcEditorSubCmd;
static NmcEditorSubCmd
parse_editor_sub_cmd (const char *cmd, char **cmd_arg)
{
NmcEditorSubCmd editor_cmd = NMC_EDITOR_SUB_CMD_UNKNOWN;
char **vec;
cli: fix an 'nmcli con edit' crash nmcli> desc AAA
2013-09-11 18:05:32 +02:00
vec = nmc_strsplit_set (cmd, " \t", 2);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
if (g_strv_length (vec) < 1) {
if (cmd_arg)
*cmd_arg = NULL;
return NMC_EDITOR_SUB_CMD_UNKNOWN;
}
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (vec[0], "set"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_SET;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "add"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_ADD;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "change"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_CHANGE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "remove"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_REMOVE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "describe"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_DESCRIBE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "print"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_PRINT;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "back"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_BACK;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "help") || strcmp (vec[0], "?") == 0)
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_HELP;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (vec[0], "quit"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_cmd = NMC_EDITOR_SUB_CMD_QUIT;
/* set pointer to command argument */
if (cmd_arg)
*cmd_arg = g_strdup (vec[1]);
g_strfreev (vec);
return editor_cmd;
}
static void
editor_sub_help (void)
{
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("------------------------------------------------------------------------------\n");
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
/* TRANSLATORS: do not translate command names and keywords before ::
* However, you should translate terms enclosed in <>.
*/
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("---[ Property menu ]---\n"
"set [<value>] :: set new value\n"
"add [<value>] :: add new option to the property\n"
"change :: change current value\n"
"remove [<index> | <option>] :: delete the value\n"
"describe :: describe property\n"
"print [setting | connection] :: print property (setting/connection) value(s)\n"
"back :: go to upper level\n"
"help/? [<command>] :: print this help or command description\n"
"quit :: exit nmcli\n"));
g_print ("------------------------------------------------------------------------------\n");
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
}
static void
editor_sub_usage (const char *command)
{
if (!command)
editor_sub_help ();
else {
/* detailed command descriptions */
NmcEditorSubCmd cmdsub = parse_editor_sub_cmd (command, NULL);
switch (cmdsub) {
case NMC_EDITOR_SUB_CMD_SET:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("set [<value>] :: set new value\n\n"
"This command sets provided <value> to this property\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_ADD:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("add [<value>] :: append new value to the property\n\n"
"This command adds provided <value> to this property, if "
"the property is of a container type. For single-valued "
"properties the property value is replaced (same as 'set').\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_CHANGE:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("change :: change current value\n\n"
"Displays current value and allows editing it.\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_REMOVE:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("remove [<value>|<index>|<option name>] :: delete the value\n\n"
"Removes the property value. For single-valued properties, this sets the\n"
"property back to its default value. For container-type properties, this removes\n"
"all the values of that property, or you can specify an argument to remove just\n"
"a single item or option. The argument is either a value or index of the item to\n"
"remove, or an option name (for properties with named options).\n\n"
"Examples: nmcli ipv4.dns> remove 8.8.8.8\n"
" nmcli ipv4.dns> remove 2\n"
" nmcli bond.options> remove downdelay\n\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_DESCRIBE:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("describe :: describe property\n\n"
"Shows property description. You can consult nm-settings(5) "
"manual page to see all NM settings and properties.\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_PRINT:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("print [property|setting|connection] :: print property (setting, connection) value(s)\n\n"
"Shows property value. Providing an argument you can also display "
"values for the whole setting or connection.\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_BACK:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("back :: go to upper menu level\n\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_HELP:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("help/? [<command>] :: help for nmcli commands\n\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_QUIT:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("quit :: exit nmcli\n\n"
"This command exits nmcli. When the connection being edited "
"is not saved, the user is asked to confirm the action.\n"));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
default:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Unknown command: '%s'\n"), command);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
}
}
}
/*----------------------------------------------------------------------------*/
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
typedef struct {
NMDevice *device;
NMActiveConnection *ac;
guint monitor_id;
} MonitorACInfo;
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
static gboolean nmc_editor_cb_called;
static GError *nmc_editor_error;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
static MonitorACInfo *nmc_editor_monitor_ac;
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
/*
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
* Store 'error' to shared 'nmc_editor_error' and monitoring info to
* 'nmc_editor_monitor_ac' and signal the condition so that
* the 'editor-thread' thread could process that.
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
*/
static void
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
set_info_and_signal_editor_thread (GError *error, MonitorACInfo *monitor_ac_info)
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
{
nmc_editor_cb_called = TRUE;
nmc_editor_error = error ? g_error_copy (error) : NULL;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
nmc_editor_monitor_ac = monitor_ac_info;
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
}
static void
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
add_connection_editor_cb (GObject *client,
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
GAsyncResult *result,
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
gpointer user_data)
{
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
NMRemoteConnection *connection;
GError *error = NULL;
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
connection = nm_client_add_connection_finish (NM_CLIENT (client), result, &error);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
set_info_and_signal_editor_thread (error, NULL);
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
g_clear_object (&connection);
g_clear_error (&error);
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
}
static void
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
update_connection_editor_cb (GObject *connection,
GAsyncResult *result,
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
gpointer user_data)
{
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
GError *error = NULL;
nm_remote_connection_commit_changes_finish (NM_REMOTE_CONNECTION (connection),
result, &error);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
set_info_and_signal_editor_thread (error, NULL);
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
g_clear_error (&error);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
}
static gboolean
progress_activation_editor_cb (gpointer user_data)
{
MonitorACInfo *info = (MonitorACInfo *) user_data;
NMDevice *device = info->device;
NMActiveConnection *ac = info->ac;
NMActiveConnectionState ac_state;
NMDeviceState dev_state;
if (!device || !ac)
nmcli: fix an error when showing progress of activation in editor (process:7213): CRITICAL **: nm_active_connectiuon_get_state: assertion `NM_IS_ACTIVE_CONNECTION (connection)' failed
2014-03-31 17:33:52 +02:00
goto finish;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
ac_state = nm_active_connection_get_state (ac);
dev_state = nm_device_get_state (device);
nmc_terminal_show_progress (nmc_device_state_to_string (dev_state));
if ( ac_state == NM_ACTIVE_CONNECTION_STATE_ACTIVATED
|| dev_state == NM_DEVICE_STATE_ACTIVATED) {
nmc_terminal_erase_line ();
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Connection successfully activated (D-Bus active path: %s)\n"),
nm_object_get_path (NM_OBJECT (ac)));
nmcli: fix an error when showing progress of activation in editor (process:7213): CRITICAL **: nm_active_connectiuon_get_state: assertion `NM_IS_ACTIVE_CONNECTION (connection)' failed
2014-03-31 17:33:52 +02:00
goto finish; /* we are done */
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
} else if ( ac_state == NM_ACTIVE_CONNECTION_STATE_DEACTIVATED
nmcli: fix an error when showing progress of activation in editor (process:7213): CRITICAL **: nm_active_connectiuon_get_state: assertion `NM_IS_ACTIVE_CONNECTION (connection)' failed
2014-03-31 17:33:52 +02:00
|| dev_state == NM_DEVICE_STATE_FAILED) {
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
nmc_terminal_erase_line ();
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: Connection activation failed.\n"));
nmcli: fix an error when showing progress of activation in editor (process:7213): CRITICAL **: nm_active_connectiuon_get_state: assertion `NM_IS_ACTIVE_CONNECTION (connection)' failed
2014-03-31 17:33:52 +02:00
goto finish; /* we are done */
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
}
return TRUE;
nmcli: fix an error when showing progress of activation in editor (process:7213): CRITICAL **: nm_active_connectiuon_get_state: assertion `NM_IS_ACTIVE_CONNECTION (connection)' failed
2014-03-31 17:33:52 +02:00
finish:
cli: clear source when progress callback terminates
2016-09-02 21:37:37 +02:00
info->monitor_id = 0;
nmcli: fix an error when showing progress of activation in editor (process:7213): CRITICAL **: nm_active_connectiuon_get_state: assertion `NM_IS_ACTIVE_CONNECTION (connection)' failed
2014-03-31 17:33:52 +02:00
if (device)
g_object_unref (device);
if (ac)
g_object_unref (ac);
return FALSE;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
}
static void
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
activate_connection_editor_cb (GObject *client,
GAsyncResult *result,
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
gpointer user_data)
{
ActivateConnectionInfo *info = (ActivateConnectionInfo *) user_data;
NMDevice *device = info->device;
const GPtrArray *ac_devs;
MonitorACInfo *monitor_ac_info = NULL;
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
NMActiveConnection *active;
GError *error = NULL;
active = nm_client_activate_connection_finish (NM_CLIENT (client), result, &error);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
if (!error) {
if (!device) {
ac_devs = nm_active_connection_get_devices (active);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
device = ac_devs->len > 0 ? g_ptr_array_index (ac_devs, 0) : NULL;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
}
if (device) {
cli: fix wrong memory allocation size Fixes: 8d9718bd0f66c74a25ad0ef9fdecbebec749b0d2
2016-09-02 21:14:30 +02:00
monitor_ac_info = g_malloc0 (sizeof (MonitorACInfo));
nmcli: fix an error when showing progress of activation in editor (process:7213): CRITICAL **: nm_active_connectiuon_get_state: assertion `NM_IS_ACTIVE_CONNECTION (connection)' failed
2014-03-31 17:33:52 +02:00
monitor_ac_info->device = g_object_ref (device);
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
monitor_ac_info->ac = active;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
monitor_ac_info->monitor_id = g_timeout_add (120, progress_activation_editor_cb, monitor_ac_info);
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
} else
g_object_unref (active);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
}
set_info_and_signal_editor_thread (error, monitor_ac_info);
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
g_clear_error (&error);
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
}
/*----------------------------------------------------------------------------*/
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
static void
print_property_description (NMSetting *setting, const char *prop_name)
{
char *desc;
desc = nmc_setting_get_property_desc (setting, prop_name);
cli: do not show description of unsupported properties Properties without meta data cannot be set or displayed: do not show them.
2017-04-12 09:17:51 +02:00
if (desc) {
g_print ("\n=== [%s] ===\n%s\n", prop_name, desc);
g_free (desc);
}
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
}
static void
print_setting_description (NMSetting *setting)
{
/* Show description of all properties */
char **all_props;
int i;
all_props = nmc_setting_get_valid_properties (setting);
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (("<<< %s >>>\n"), nm_setting_get_name (setting));
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
for (i = 0; all_props && all_props[i]; i++)
print_property_description (setting, all_props[i]);
g_strfreev (all_props);
}
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
static gboolean
connection_remove_setting (NMConnection *connection, NMSetting *setting)
{
gboolean mandatory;
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
g_return_val_if_fail (setting, FALSE);
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
mandatory = is_setting_mandatory (connection, setting);
if (!mandatory) {
nm_connection_remove_setting (connection, G_OBJECT_TYPE (setting));
return TRUE;
}
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: setting '%s' is mandatory and cannot be removed.\n"),
nm_setting_get_name (setting));
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
return FALSE;
}
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
static void
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
editor_show_status_line (NMConnection *connection, gboolean dirty, gboolean temp)
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
{
NMSettingConnection *s_con;
const char *con_type, *con_id, *con_uuid;
s_con = nm_connection_get_setting_connection (connection);
g_assert (s_con);
con_type = nm_setting_connection_get_connection_type (s_con);
con_id = nm_connection_get_id (connection);
con_uuid = nm_connection_get_uuid (connection);
/* TRANSLATORS: status line in nmcli connection editor */
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("[ Type: %s | Name: %s | UUID: %s | Dirty: %s | Temp: %s ]\n"),
con_type, con_id, con_uuid,
dirty ? _("yes") : _("no"),
temp ? _("yes") : _("no"));
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
}
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
static gboolean
refresh_remote_connection (GWeakRef *weak, NMRemoteConnection **remote)
{
gboolean previous;
g_return_val_if_fail (remote != NULL, FALSE);
previous = (*remote != NULL);
if (*remote)
g_object_unref (*remote);
*remote = g_weak_ref_get (weak);
return (previous && !*remote);
}
cli: always get up-to-date dirty status
2014-01-20 18:06:37 +01:00
static gboolean
is_connection_dirty (NMConnection *connection, NMRemoteConnection *remote)
{
return !nm_connection_compare (connection,
remote ? NM_CONNECTION (remote) : NULL,
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
NM_SETTING_COMPARE_FLAG_IGNORE_SECRETS |
cli: ignore timestamp when comparing connections in the editor
2014-11-18 14:24:26 +01:00
NM_SETTING_COMPARE_FLAG_IGNORE_TIMESTAMP);
cli: always get up-to-date dirty status
2014-01-20 18:06:37 +01:00
}
cli: use readline for quit confirmation in editor, and accept localized answer
2014-06-02 12:24:56 +02:00
static gboolean
confirm_quit (void)
{
char *answer;
gboolean want_quit = FALSE;
answer = nmc_readline (_("The connection is not saved. "
"Do you really want to quit? %s"),
prompt_yes_no (FALSE, NULL));
answer = answer ? g_strstrip (answer) : NULL;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (answer && matches (answer, WORD_LOC_YES))
cli: use readline for quit confirmation in editor, and accept localized answer
2014-06-02 12:24:56 +02:00
want_quit = TRUE;
g_free (answer);
return want_quit;
}
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
/*
Fix typos https://bugzilla.gnome.org/show_bug.cgi?id=710505
2013-10-19 17:38:00 +03:00
* Submenu for detailed property editing
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
* Return: TRUE - continue; FALSE - should quit
*/
static gboolean
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
property_edit_submenu (NmCli *nmc,
NMConnection *connection,
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
NMRemoteConnection **rem_con,
GWeakRef *rem_con_weak,
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
NMSetting *curr_setting,
const char *prop_name)
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
{
NmcEditorSubCmd cmdsub;
gboolean cmd_property_loop = TRUE;
gboolean should_quit = FALSE;
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
char *prop_val_user;
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
gboolean set_result;
GError *tmp_err = NULL;
char *prompt;
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
gboolean dirty;
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
GValue prop_g_value = G_VALUE_INIT;
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
gboolean temp_changes;
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
gboolean removed;
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
/* Set global variable for use in TAB completion */
nmc_tab_completion.property = prop_name;
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
prompt = nmc_colorize (nmc->nmc_config.use_colors, nmc->editor_prompt_color, NM_META_TERM_FORMAT_NORMAL,
nmcli: add support for terminal formatting, like bold, dim, underline, etc.
2015-02-17 13:20:37 +01:00
"nmcli %s.%s> ",
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
nm_setting_get_name (curr_setting), prop_name);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
while (cmd_property_loop) {
char *cmd_property_user;
char *cmd_property_arg;
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
/* Get the remote connection again, it may have disapeared */
removed = refresh_remote_connection (rem_con_weak, rem_con);
if (removed)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("The connection profile has been removed from another client. "
"You may type 'save' in the main menu to restore it.\n"));
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
/* Connection is dirty? (not saved or differs from the saved) */
cli: always get up-to-date dirty status
2014-01-20 18:06:37 +01:00
dirty = is_connection_dirty (connection, *rem_con);
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
temp_changes = *rem_con ? nm_remote_connection_get_unsaved (*rem_con) : TRUE;
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
if (nmc->editor_status_line)
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
editor_show_status_line (connection, dirty, temp_changes);
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
cmd_property_user = nmc_readline ("%s", prompt);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
if (!cmd_property_user || *cmd_property_user == '\0')
continue;
cmdsub = parse_editor_sub_cmd (g_strstrip (cmd_property_user), &cmd_property_arg);
switch (cmdsub) {
case NMC_EDITOR_SUB_CMD_SET:
case NMC_EDITOR_SUB_CMD_ADD:
/* list, arrays,...: SET replaces the whole property value
* ADD adds the new value(s)
* single values: : both SET and ADD sets the new value
*/
cli: print valid-values hint in property menu for empty "set" command nmcli 802-11-wireless.mode> set Allowed values for 'mode' property: infrastructure, adhoc, ap Enter 'mode' value:
2015-05-11 17:07:20 +02:00
if (!cmd_property_arg) {
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_strfreev char **to_free = NULL;
const char *const*avals;
cli: add values_func() to NmcPropertyInfo and use for "connection" setting
2017-03-25 13:20:11 +01:00
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
avals = nmc_setting_get_property_allowed_values (curr_setting, prop_name, &to_free);
cli: print valid-values hint in property menu for empty "set" command nmcli 802-11-wireless.mode> set Allowed values for 'mode' property: infrastructure, adhoc, ap Enter 'mode' value:
2015-05-11 17:07:20 +02:00
if (avals) {
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_free char *avals_str = NULL;
avals_str = nmc_util_strv_for_display (avals, FALSE);
cli: print valid-values hint in property menu for empty "set" command nmcli 802-11-wireless.mode> set Allowed values for 'mode' property: infrastructure, adhoc, ap Enter 'mode' value:
2015-05-11 17:07:20 +02:00
g_print (_("Allowed values for '%s' property: %s\n"),
prop_name, avals_str);
}
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
prop_val_user = nmc_readline (_("Enter '%s' value: "), prop_name);
cli: print valid-values hint in property menu for empty "set" command nmcli 802-11-wireless.mode> set Allowed values for 'mode' property: infrastructure, adhoc, ap Enter 'mode' value:
2015-05-11 17:07:20 +02:00
} else
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
prop_val_user = g_strdup (cmd_property_arg);
/* nmc_setting_set_property() only adds new value, thus we have to
* remove the original value and save it for error cases.
*/
if (cmdsub == NMC_EDITOR_SUB_CMD_SET) {
nmc_property_get_gvalue (curr_setting, prop_name, &prop_g_value);
nmc_property_set_default_value (curr_setting, prop_name);
}
set_result = nmc_setting_set_property (curr_setting, prop_name, prop_val_user, &tmp_err);
g_free (prop_val_user);
if (!set_result) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: failed to set '%s' property: %s\n"), prop_name, tmp_err->message);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
g_clear_error (&tmp_err);
cli: fix an error path in second level 'set' command of nmcli editor Test case: nmcli> goto ipv4.addresses nmcli ipv4.addresses> set 1.1.1.1 Do you also want to set 'ipv4.method' to 'manual'? [yes]: no nmcli ipv4.addresses> set 3.3.3.3/99 Error: failed to set 'addresses' property: invalid prefix '99'; <1-32> allowed Do you also want to set 'ipv4.method' to 'manual'? [yes]: The second question (after failed 'set') should not be asked.
2014-11-06 14:49:03 +01:00
if (cmdsub == NMC_EDITOR_SUB_CMD_SET) {
/* Block change signals and restore original value */
g_signal_handlers_block_matched (curr_setting, G_SIGNAL_MATCH_DATA, 0, 0, NULL, NULL, NULL);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
nmc_property_set_gvalue (curr_setting, prop_name, &prop_g_value);
cli: fix an error path in second level 'set' command of nmcli editor Test case: nmcli> goto ipv4.addresses nmcli ipv4.addresses> set 1.1.1.1 Do you also want to set 'ipv4.method' to 'manual'? [yes]: no nmcli ipv4.addresses> set 3.3.3.3/99 Error: failed to set 'addresses' property: invalid prefix '99'; <1-32> allowed Do you also want to set 'ipv4.method' to 'manual'? [yes]: The second question (after failed 'set') should not be asked.
2014-11-06 14:49:03 +01:00
g_signal_handlers_unblock_matched (curr_setting, G_SIGNAL_MATCH_DATA, 0, 0, NULL, NULL, NULL);
}
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
}
if (G_IS_VALUE (&prop_g_value))
g_value_unset (&prop_g_value);
break;
case NMC_EDITOR_SUB_CMD_CHANGE:
cli: add support for secret agent to 'nmcli dev connect' too And move secrets getting code to common.c (without changes).
2014-11-20 12:36:14 +01:00
rl_startup_hook = nmc_rl_set_deftext;
cli: refactor property to string conversion Instead of having a get_func() and out2in_func(), have only one get_func() that accepts an argument of the output format. This way, a conversion to parsable input format, doesn't have to go first thourgh get_func() and mangle the pretty string in out2in_func(). This fixes conversions via nmc_property_out2in_cut_paren(). For example, nmc_property_802_1X_get_private_key_password_flags() would return a localized string _("0 (none)"). There is no guarantee that out2in_func() would find the expected output format after localizing. This also fixes nmc_property_out2in_routes() which expected a format "dst =" (would be "ip =") and expects mandatory 'nh' and 'mt' arguments. In fact, the regex didn't match and nmc_property_out2in_routes() always failed. While at it, also combine the implementation of nmc_property_ipv4_get_routes() and nmc_property_ipv6_get_routes().
2015-05-21 13:00:40 +02:00
nmc_rl_pre_input_deftext = nmc_setting_get_property_parsable (curr_setting, prop_name, NULL);
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
prop_val_user = nmc_readline (_("Edit '%s' value: "), prop_name);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
nmc_property_get_gvalue (curr_setting, prop_name, &prop_g_value);
nmc_property_set_default_value (curr_setting, prop_name);
if (!nmc_setting_set_property (curr_setting, prop_name, prop_val_user, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: failed to set '%s' property: %s\n"), prop_name, tmp_err->message);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
g_clear_error (&tmp_err);
cli: fix an error path in second level 'set' command of nmcli editor Test case: nmcli> goto ipv4.addresses nmcli ipv4.addresses> set 1.1.1.1 Do you also want to set 'ipv4.method' to 'manual'? [yes]: no nmcli ipv4.addresses> set 3.3.3.3/99 Error: failed to set 'addresses' property: invalid prefix '99'; <1-32> allowed Do you also want to set 'ipv4.method' to 'manual'? [yes]: The second question (after failed 'set') should not be asked.
2014-11-06 14:49:03 +01:00
g_signal_handlers_block_matched (curr_setting, G_SIGNAL_MATCH_DATA, 0, 0, NULL, NULL, NULL);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
nmc_property_set_gvalue (curr_setting, prop_name, &prop_g_value);
cli: fix an error path in second level 'set' command of nmcli editor Test case: nmcli> goto ipv4.addresses nmcli ipv4.addresses> set 1.1.1.1 Do you also want to set 'ipv4.method' to 'manual'? [yes]: no nmcli ipv4.addresses> set 3.3.3.3/99 Error: failed to set 'addresses' property: invalid prefix '99'; <1-32> allowed Do you also want to set 'ipv4.method' to 'manual'? [yes]: The second question (after failed 'set') should not be asked.
2014-11-06 14:49:03 +01:00
g_signal_handlers_unblock_matched (curr_setting, G_SIGNAL_MATCH_DATA, 0, 0, NULL, NULL, NULL);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
}
g_free (prop_val_user);
if (G_IS_VALUE (&prop_g_value))
g_value_unset (&prop_g_value);
break;
case NMC_EDITOR_SUB_CMD_REMOVE:
if (cmd_property_arg) {
unsigned long val_int = G_MAXUINT32;
char *option = NULL;
if (!nmc_string_to_uint (cmd_property_arg, TRUE, 0, G_MAXUINT32, &val_int))
option = g_strdup (cmd_property_arg);
if (!nmc_setting_remove_property_option (curr_setting, prop_name,
option ? g_strstrip (option) : NULL,
(guint32) val_int,
&tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: %s\n"), tmp_err->message);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
g_clear_error (&tmp_err);
}
g_free (option);
cli: deny removing values of nmcli-unchangable properties nmc_setting_reset_property() function checks whether we allow changing the property (set_func != NULL) and if so, the property value is reset to default.
2013-09-06 14:56:41 +02:00
} else {
if (!nmc_setting_reset_property (curr_setting, prop_name, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: failed to remove value of '%s': %s\n"), prop_name,
tmp_err->message);
cli: deny removing values of nmcli-unchangable properties nmc_setting_reset_property() function checks whether we allow changing the property (set_func != NULL) and if so, the property value is reset to default.
2013-09-06 14:56:41 +02:00
g_clear_error (&tmp_err);
}
}
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
case NMC_EDITOR_SUB_CMD_DESCRIBE:
/* Show property description */
print_property_description (curr_setting, prop_name);
break;
case NMC_EDITOR_SUB_CMD_PRINT:
/* Print current connection settings/properties */
if (cmd_property_arg) {
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (cmd_property_arg, "setting"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_show_setting (curr_setting, nmc);
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if ( matches (cmd_property_arg, "connection")
|| matches (cmd_property_arg, "all"))
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
editor_show_connection (connection, nmc);
else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Unknown command argument: '%s'\n"), cmd_property_arg);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
} else {
cli: editor 'change' command: output -> input format conversion (rh #998929) The format of property values that nmcli prints is not the same (for some properties) as the format nmcli editor accepts as input from user. The reasons are that (a) output format is more descriptive and not much suitable to be typed, (b) it comes in most cases from libnm-util. 'change' command displays current property value and allows users to edit it. So we convert the output into input format, before presenting it to the user. https://bugzilla.redhat.com/show_bug.cgi?id=998929
2013-09-11 17:32:00 +02:00
char *prop_val = nmc_setting_get_property (curr_setting, prop_name, NULL);
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("%s: %s\n", prop_name, prop_val);
cli: editor 'change' command: output -> input format conversion (rh #998929) The format of property values that nmcli prints is not the same (for some properties) as the format nmcli editor accepts as input from user. The reasons are that (a) output format is more descriptive and not much suitable to be typed, (b) it comes in most cases from libnm-util. 'change' command displays current property value and allows users to edit it. So we convert the output into input format, before presenting it to the user. https://bugzilla.redhat.com/show_bug.cgi?id=998929
2013-09-11 17:32:00 +02:00
g_free (prop_val);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
}
break;
case NMC_EDITOR_SUB_CMD_BACK:
cli: TAB-completion for enum-style property values (rh #1034126) Valid values for enumeration-style properties are offered in TAB-completion in the editor. Thus the user has a quick overview of the possible values and can edit properties more easily. Example: $ nmcli con edit type wifi nmcli> set wifi-sec.group <TAB> ccmp tkip wep104 wep40 nmcli> ... https://bugzilla.redhat.com/show_bug.cgi?id=1034126
2015-04-22 10:47:06 +02:00
/* Set global variable for use in TAB completion */
nmc_tab_completion.property = NULL;
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
cmd_property_loop = FALSE;
break;
case NMC_EDITOR_SUB_CMD_HELP:
editor_sub_usage (cmd_property_arg);
break;
case NMC_EDITOR_SUB_CMD_QUIT:
cli: always get up-to-date dirty status
2014-01-20 18:06:37 +01:00
if (is_connection_dirty (connection, *rem_con)) {
cli: use readline for quit confirmation in editor, and accept localized answer
2014-06-02 12:24:56 +02:00
if (confirm_quit ()) {
cli: use 'dirty' flag for "Do you really want to quit?" confirmation in editor
2013-04-23 12:53:52 +02:00
cmd_property_loop = FALSE;
should_quit = TRUE; /* we will quit nmcli */
}
} else {
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
cmd_property_loop = FALSE;
should_quit = TRUE; /* we will quit nmcli */
}
break;
case NMC_EDITOR_SUB_CMD_UNKNOWN:
default:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Unknown command: '%s'\n"), cmd_property_user);
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
break;
}
g_free (cmd_property_user);
g_free (cmd_property_arg);
}
g_free (prompt);
return !should_quit;
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/*
* Split 'str' in the following format: [[[setting.]property] [value]]
* and return the components in 'setting', 'property' and 'value'
* Use g_free() to deallocate the returned strings.
*/
static void
split_editor_main_cmd_args (const char *str, char **setting, char **property, char **value)
{
char **args, **items;
if (!str)
return;
cli: fix an 'nmcli con edit' crash nmcli> desc AAA
2013-09-11 18:05:32 +02:00
args = nmc_strsplit_set (str, " \t", 2);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (args[0]) {
cli: fix an 'nmcli con edit' crash nmcli> desc AAA
2013-09-11 18:05:32 +02:00
items = nmc_strsplit_set (args[0], ".", 2);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (g_strv_length (items) == 2) {
if (setting)
*setting = g_strdup (items[0]);
if (property)
*property = g_strdup (items[1]);
} else {
if (property)
*property = g_strdup (items[0]);
}
g_strfreev (items);
if (value && args[1])
cli: fix an 'nmcli con edit' crash nmcli> desc AAA
2013-09-11 18:05:32 +02:00
*value = g_strstrip (g_strdup (args[1]));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
g_strfreev (args);
}
static NMSetting *
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
create_setting_by_name (const char *name, const NameItem *valid_settings_main, const NameItem *valid_settings_slave)
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{
const char *setting_name;
NMSetting *setting = NULL;
/* Get a valid setting name */
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
setting_name = check_valid_name (name, valid_settings_main, valid_settings_slave, NULL);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (setting_name) {
cli: simplify setting creation
2013-06-05 16:04:06 -05:00
setting = nmc_setting_new_for_name (setting_name);
if (!setting)
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
return NULL; /* This should really not happen */
nmc_setting_custom_init (setting);
}
return setting;
}
static const char *
ask_check_setting (const char *arg,
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const NameItem *valid_settings_main,
const NameItem *valid_settings_slave,
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
const char *valid_settings_str)
{
char *setting_name_user;
const char *setting_name;
GError *err = NULL;
if (!arg) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Available settings: %s\n"), valid_settings_str);
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
setting_name_user = nmc_readline (EDITOR_PROMPT_SETTING);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
} else
setting_name_user = g_strdup (arg);
if (setting_name_user)
g_strstrip (setting_name_user);
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
if (!(setting_name = check_valid_name (setting_name_user,
valid_settings_main,
valid_settings_slave,
&err))) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid setting name; %s\n"), err->message);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&err);
}
g_free (setting_name_user);
return setting_name;
}
static const char *
ask_check_property (const char *arg,
const char **valid_props,
const char *valid_props_str)
{
char *prop_name_user;
const char *prop_name;
GError *tmp_err = NULL;
if (!arg) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Available properties: %s\n"), valid_props_str);
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
prop_name_user = nmc_readline (EDITOR_PROMPT_PROPERTY);
nmcli: fix segfault in edit mode when parsing user prompt readline_x returns NULL for empty input, so it is wrong to call g_strstrip without checking for NULL first. Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-09-05 13:59:51 +02:00
if (prop_name_user)
g_strstrip (prop_name_user);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
} else
prop_name_user = g_strdup (arg);
if (!(prop_name = nmc_string_is_valid (prop_name_user, valid_props, &tmp_err))) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: property %s\n"), tmp_err->message);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&tmp_err);
}
g_free (prop_name_user);
return prop_name;
}
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
/* Copy timestamp from src do dst */
static void
update_connection_timestamp (NMConnection *src, NMConnection *dst)
{
NMSettingConnection *s_con_src, *s_con_dst;
s_con_src = nm_connection_get_setting_connection (src);
s_con_dst = nm_connection_get_setting_connection (dst);
if (s_con_src && s_con_dst) {
guint64 timestamp = nm_setting_connection_get_timestamp (s_con_src);
g_object_set (s_con_dst, NM_SETTING_CONNECTION_TIMESTAMP, timestamp, NULL);
}
}
cli: confirm connection saving when autoconnect=yes (rh #953296) Ask for confirmation in these cases: * autoconnect=yes and connection has not been saved yet * autoconnect=yes, but saved connection had autoconnect=no The confirmation can be switched off with nmcli> nmcli save-confirmation no
2013-07-23 14:25:52 +02:00
static gboolean
confirm_connection_saving (NMConnection *local, NMConnection *remote)
{
NMSettingConnection *s_con_loc, *s_con_rem;
gboolean ac_local, ac_remote;
gboolean confirmed = TRUE;
s_con_loc = nm_connection_get_setting_connection (local);
g_assert (s_con_loc);
ac_local = nm_setting_connection_get_autoconnect (s_con_loc);
if (remote) {
s_con_rem = nm_connection_get_setting_connection (remote);
g_assert (s_con_rem);
ac_remote = nm_setting_connection_get_autoconnect (s_con_rem);
} else
ac_remote = FALSE;
if (ac_local && !ac_remote) {
char *answer;
cli: implement TAB completion for inputs in questionnaire mode and also - accept yes/no options as localized strings - simplify some questionnaire code
2014-05-14 16:42:26 +02:00
answer = nmc_readline (_("Saving the connection with 'autoconnect=yes'. "
"That might result in an immediate activation of the connection.\n"
"Do you still want to save? %s"), prompt_yes_no (TRUE, NULL));
answer = answer ? g_strstrip (answer) : NULL;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (!answer || matches (answer, WORD_LOC_YES))
cli: confirm connection saving when autoconnect=yes (rh #953296) Ask for confirmation in these cases: * autoconnect=yes and connection has not been saved yet * autoconnect=yes, but saved connection had autoconnect=no The confirmation can be switched off with nmcli> nmcli save-confirmation no
2013-07-23 14:25:52 +02:00
confirmed = TRUE;
else
confirmed = FALSE;
g_free (answer);
}
return confirmed;
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
typedef struct {
guint level;
char *main_prompt;
NMSetting *curr_setting;
char **valid_props;
char *valid_props_str;
} NmcEditorMenuContext;
static void
cli: pass use_colors as separate option instead of global nmc nmc contains all options and collects output data. It is very hard to understand what it does. Start splitting it up, and pass arguments along -- as needed.
2017-03-30 12:23:26 +02:00
menu_switch_to_level0 (NmcColorOption color_option,
cli: move the decision whether to use colors to nmc_colorize() This allows us to use a conditionally colorized output outside print_required_fields().
2015-03-26 17:51:13 +01:00
NmcEditorMenuContext *menu_ctx,
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
const char *prompt,
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
NMMetaTermColor prompt_color)
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{
menu_ctx->level = 0;
g_free (menu_ctx->main_prompt);
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
menu_ctx->main_prompt = nmc_colorize (color_option, prompt_color, NM_META_TERM_FORMAT_NORMAL, "%s", prompt);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
menu_ctx->curr_setting = NULL;
g_strfreev (menu_ctx->valid_props);
menu_ctx->valid_props = NULL;
g_free (menu_ctx->valid_props_str);
menu_ctx->valid_props_str = NULL;
}
static void
cli: pass use_colors as separate option instead of global nmc nmc contains all options and collects output data. It is very hard to understand what it does. Start splitting it up, and pass arguments along -- as needed.
2017-03-30 12:23:26 +02:00
menu_switch_to_level1 (NmcColorOption color_option,
cli: move the decision whether to use colors to nmc_colorize() This allows us to use a conditionally colorized output outside print_required_fields().
2015-03-26 17:51:13 +01:00
NmcEditorMenuContext *menu_ctx,
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NMSetting *setting,
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
const char *setting_name,
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
NMMetaTermColor prompt_color)
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{
menu_ctx->level = 1;
g_free (menu_ctx->main_prompt);
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
menu_ctx->main_prompt = nmc_colorize (color_option, prompt_color, NM_META_TERM_FORMAT_NORMAL,
nmcli: add support for terminal formatting, like bold, dim, underline, etc.
2015-02-17 13:20:37 +01:00
"nmcli %s> ", setting_name);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
menu_ctx->curr_setting = setting;
g_strfreev (menu_ctx->valid_props);
menu_ctx->valid_props = nmc_setting_get_valid_properties (menu_ctx->curr_setting);
g_free (menu_ctx->valid_props_str);
menu_ctx->valid_props_str = g_strjoinv (", ", menu_ctx->valid_props);
}
static gboolean
editor_menu_main (NmCli *nmc, NMConnection *connection, const char *connection_type)
{
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
NMSettingConnection *s_con;
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
NMRemoteConnection *rem_con;
NMRemoteConnection *con_tmp;
GWeakRef weak = { { NULL } };
gboolean removed;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NmcEditorMainCmd cmd;
char *cmd_user;
gboolean cmd_loop = TRUE;
char *cmd_arg = NULL;
char *cmd_arg_s, *cmd_arg_p, *cmd_arg_v;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const NameItem *valid_settings_main = NULL;
const NameItem *valid_settings_slave = NULL;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
char *valid_settings_str = NULL;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
const char *s_type = NULL;
char *slv_type;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
AddConnectionInfo *info = NULL;
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
gboolean dirty;
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
gboolean temp_changes;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
GError *err1 = NULL;
NmcEditorMenuContext menu_ctx;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
s_con = nm_connection_get_setting_connection (connection);
if (s_con)
s_type = nm_setting_connection_get_slave_type (s_con);
slv_type = g_strdup_printf ("%s-slave", s_type ? s_type : "no");
valid_settings_main = get_valid_settings_array (connection_type);
valid_settings_slave = get_valid_settings_array (slv_type);
g_free (slv_type);
valid_settings_str = get_valid_options_string (valid_settings_main, valid_settings_slave);
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("You may edit the following settings: %s\n"), valid_settings_str);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
menu_ctx.level = 0;
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
menu_ctx.main_prompt = nmc_colorize (nmc->nmc_config.use_colors, nmc->editor_prompt_color, NM_META_TERM_FORMAT_NORMAL,
nmcli: add support for terminal formatting, like bold, dim, underline, etc.
2015-02-17 13:20:37 +01:00
BASE_PROMPT);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
menu_ctx.curr_setting = NULL;
menu_ctx.valid_props = NULL;
menu_ctx.valid_props_str = NULL;
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
/* Get remote connection */
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
con_tmp = nm_client_get_connection_by_uuid (nmc->client,
nm_connection_get_uuid (connection));
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
g_weak_ref_init (&weak, con_tmp);
rem_con = g_weak_ref_get (&weak);
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
while (cmd_loop) {
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
/* Connection is dirty? (not saved or differs from the saved) */
cli: always get up-to-date dirty status
2014-01-20 18:06:37 +01:00
dirty = is_connection_dirty (connection, rem_con);
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
temp_changes = rem_con ? nm_remote_connection_get_unsaved (rem_con) : TRUE;
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
if (nmc->editor_status_line)
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
editor_show_status_line (connection, dirty, temp_changes);
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
/* Read user input */
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
cmd_user = nmc_readline ("%s", menu_ctx.main_prompt);
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
/* Get the remote connection again, it may have disapeared */
removed = refresh_remote_connection (&weak, &rem_con);
if (removed)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("The connection profile has been removed from another client. "
"You may type 'save' to restore it.\n"));
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
cli: line-editing capability for interactive editor using an edit line library Supported libraries are: GNU Readline (libreadline) http://cnswww.cns.cwru.edu/php/chet/readline/rltop.html NetBSD Editline (libedit) http://www.thrysoee.dk/editline/ We load these symbols: readline() - main function allowing getting user input and aditing that add_history() - stores strings to history rl_insert_text() - insert text into the line at the current cursor position rl_startup_hook - hook to call just before readline prints the first prompt
2013-01-31 19:03:51 +01:00
if (!cmd_user || *cmd_user == '\0')
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
continue;
cmd = parse_editor_main_cmd (g_strstrip (cmd_user), &cmd_arg);
cmd_arg_s = NULL;
cmd_arg_p = NULL;
cmd_arg_v = NULL;
split_editor_main_cmd_args (cmd_arg, &cmd_arg_s, &cmd_arg_p, &cmd_arg_v);
switch (cmd) {
case NMC_EDITOR_MAIN_CMD_SET:
/* Set property value */
if (!cmd_arg) {
if (menu_ctx.level == 1) {
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_strfreev char **avals_to_free = NULL;
gs_free char *prop_val_user = NULL;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
const char *prop_name;
cli: add values_func() to NmcPropertyInfo and use for "connection" setting
2017-03-25 13:20:11 +01:00
const char *const*avals;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
GError *tmp_err = NULL;
prop_name = ask_check_property (cmd_arg,
(const char **) menu_ctx.valid_props,
menu_ctx.valid_props_str);
if (!prop_name)
break;
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
avals = nmc_setting_get_property_allowed_values (menu_ctx.curr_setting, prop_name, &avals_to_free);
cli: change function for allowed values to return array of strings
2015-04-22 12:10:53 +02:00
if (avals) {
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_free char *avals_str = NULL;
avals_str = nmc_util_strv_for_display (avals, FALSE);
cli: change function for allowed values to return array of strings
2015-04-22 12:10:53 +02:00
g_print (_("Allowed values for '%s' property: %s\n"),
prop_name, avals_str);
}
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
prop_val_user = nmc_readline (_("Enter '%s' value: "), prop_name);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* Set property value */
if (!nmc_setting_set_property (menu_ctx.curr_setting, prop_name, prop_val_user, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: failed to set '%s' property: %s\n"), prop_name, tmp_err->message);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&tmp_err);
}
} else {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: no setting selected; valid are [%s]\n"), valid_settings_str);
g_print (_("use 'goto <setting>' first, or 'set <setting>.<property>'\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
} else {
NMSetting *ss = NULL;
gboolean created_ss = FALSE;
char *prop_name;
GError *tmp_err = NULL;
if (cmd_arg_s) {
/* setting provided as "setting.property" */
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
ss = is_setting_valid (connection, valid_settings_main, valid_settings_slave, cmd_arg_s);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (!ss) {
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
ss = create_setting_by_name (cmd_arg_s, valid_settings_main, valid_settings_slave);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (!ss) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid setting argument '%s'; valid are [%s]\n"),
cmd_arg_s, valid_settings_str);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
}
created_ss = TRUE;
}
} else {
if (menu_ctx.curr_setting)
ss = menu_ctx.curr_setting;
else {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: missing setting for '%s' property\n"), cmd_arg_p);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
}
}
prop_name = is_property_valid (ss, cmd_arg_p, &tmp_err);
if (!prop_name) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid property: %s\n"), tmp_err->message);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&tmp_err);
if (created_ss)
g_object_unref (ss);
break;
}
/* Ask for value */
if (!cmd_arg_v) {
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_strfreev char **avals_to_free = NULL;
const char *const*avals;
avals = nmc_setting_get_property_allowed_values (ss, prop_name, &avals_to_free);
cli: change function for allowed values to return array of strings
2015-04-22 12:10:53 +02:00
if (avals) {
cli: allow dynamic results from values_fcn() and describe_fcn() property functions Mostly these strings are static. In same cases they are generated however. Instead of caching them in a static variable, return them to the caller. And belatedly fix invoking describe_fcn().
2017-03-27 23:11:11 +02:00
gs_free char *avals_str = NULL;
avals_str = nmc_util_strv_for_display (avals, FALSE);
cli: change function for allowed values to return array of strings
2015-04-22 12:10:53 +02:00
g_print (_("Allowed values for '%s' property: %s\n"),
prop_name, avals_str);
}
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
cmd_arg_v = nmc_readline (_("Enter '%s' value: "), prop_name);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
/* Set property value */
if (!nmc_setting_set_property (ss, prop_name, cmd_arg_v, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: failed to set '%s' property: %s\n"),
prop_name, tmp_err->message);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&tmp_err);
}
if (created_ss)
nm_connection_add_setting (connection, ss);
g_free (prop_name);
}
break;
case NMC_EDITOR_MAIN_CMD_GOTO:
/* cmd_arg_s != NULL means 'setting.property' argument */
if (menu_ctx.level == 0 || cmd_arg_s) {
/* in top level - no setting selected yet */
const char *setting_name;
NMSetting *setting;
const char *user_arg = cmd_arg_s ? cmd_arg_s : cmd_arg_p;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
setting_name = ask_check_setting (user_arg,
valid_settings_main,
valid_settings_slave,
valid_settings_str);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (!setting_name)
break;
setting = nm_connection_get_setting_by_name (connection, setting_name);
if (!setting) {
cli: simplify setting creation
2013-06-05 16:04:06 -05:00
setting = nmc_setting_new_for_name (setting_name);
if (!setting) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: unknown setting '%s'\n"), setting_name);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
}
nmc_setting_custom_init (setting);
cli: only connect handlers for property changes in interactive edit Fixes: c5324ed285aff7d6d58212e4b030a0fc556eb43b (cherry picked from commit 79c81bbb8cf5903462caf9f60e1bca4a22f3280f)
2016-08-24 15:01:03 +02:00
if (NM_IS_SETTING_WIRELESS (setting))
nmc_setting_wireless_connect_handlers (NM_SETTING_WIRELESS (setting));
else if (NM_IS_SETTING_IP4_CONFIG (setting))
nmc_setting_ip4_connect_handlers (NM_SETTING_IP_CONFIG (setting));
else if (NM_IS_SETTING_IP6_CONFIG (setting))
nmc_setting_ip6_connect_handlers (NM_SETTING_IP_CONFIG (setting));
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
else if (NM_IS_SETTING_PROXY (setting))
nmc_setting_proxy_connect_handlers (NM_SETTING_PROXY (setting));
cli: only connect handlers for property changes in interactive edit Fixes: c5324ed285aff7d6d58212e4b030a0fc556eb43b (cherry picked from commit 79c81bbb8cf5903462caf9f60e1bca4a22f3280f)
2016-08-24 15:01:03 +02:00
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
nm_connection_add_setting (connection, setting);
}
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
/* Set global variable for use in TAB completion */
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
nmc_tab_completion.setting = setting;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* Switch to level 1 */
cli: pass use_colors as separate option instead of global nmc nmc contains all options and collects output data. It is very hard to understand what it does. Start splitting it up, and pass arguments along -- as needed.
2017-03-30 12:23:26 +02:00
menu_switch_to_level1 (nmc->nmc_config.use_colors, &menu_ctx, setting, setting_name, nmc->editor_prompt_color);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (!cmd_arg_s) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("You may edit the following properties: %s\n"), menu_ctx.valid_props_str);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
}
}
if (menu_ctx.level == 1 || cmd_arg_s) {
/* level 1 - setting selected */
const char *prop_name;
prop_name = ask_check_property (cmd_arg_p,
(const char **) menu_ctx.valid_props,
menu_ctx.valid_props_str);
if (!prop_name)
break;
cli: add a submenu for connection editor for detail property handling property_edit_submenu() is the main function. It is entered from main menu by 'goto' command. Commands of this menu: set - set new value to the property add - add a value to the property (for non-container properties it is the same as set) change - allow editing the current property value remove - remove the property value(s) describe - show the property description and/or usage print - show the property value back - return to the setting level (main menu) help - show help (command descriptions) quit - quit nmcli
2013-03-04 14:07:51 +01:00
/* submenu - level 2 - editing properties */
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
cmd_loop = property_edit_submenu (nmc,
connection,
&rem_con,
&weak,
menu_ctx.curr_setting,
prop_name);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
break;
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
case NMC_EDITOR_MAIN_CMD_REMOVE:
/* Remove setting from connection, or delete value of a property */
if (!cmd_arg) {
if (menu_ctx.level == 1) {
cli: deny removing values of nmcli-unchangable properties nmc_setting_reset_property() function checks whether we allow changing the property (set_func != NULL) and if so, the property value is reset to default.
2013-09-06 14:56:41 +02:00
GError *tmp_err = NULL;
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
const char *prop_name;
prop_name = ask_check_property (cmd_arg,
(const char **) menu_ctx.valid_props,
menu_ctx.valid_props_str);
if (!prop_name)
break;
/* Delete property value */
cli: deny removing values of nmcli-unchangable properties nmc_setting_reset_property() function checks whether we allow changing the property (set_func != NULL) and if so, the property value is reset to default.
2013-09-06 14:56:41 +02:00
if (!nmc_setting_reset_property (menu_ctx.curr_setting, prop_name, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: failed to remove value of '%s': %s\n"), prop_name,
tmp_err->message);
cli: deny removing values of nmcli-unchangable properties nmc_setting_reset_property() function checks whether we allow changing the property (set_func != NULL) and if so, the property value is reset to default.
2013-09-06 14:56:41 +02:00
g_clear_error (&tmp_err);
}
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
} else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: no argument given; valid are [%s]\n"), valid_settings_str);
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
} else {
NMSetting *ss = NULL;
gboolean descr_all;
char *user_s;
/* cmd_arg_s != NULL means argument is "setting.property" */
descr_all = !cmd_arg_s && !menu_ctx.curr_setting;
user_s = descr_all ? cmd_arg_p : cmd_arg_s ? cmd_arg_s : NULL;
if (user_s) {
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
ss = is_setting_valid (connection,
valid_settings_main,
valid_settings_slave,
user_s);
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
if (!ss) {
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
if (check_valid_name (user_s,
valid_settings_main,
valid_settings_slave,
NULL))
g_print (_("Setting '%s' is not present in the connection.\n"),
user_s);
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid setting argument '%s'; valid are [%s]\n"),
user_s, valid_settings_str);
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
break;
}
} else
ss = menu_ctx.curr_setting;
if (descr_all) {
/* Remove setting from the connection */
connection_remove_setting (connection, ss);
if (ss == menu_ctx.curr_setting) {
/* If we removed the setting we are in, go up */
cli: pass use_colors as separate option instead of global nmc nmc contains all options and collects output data. It is very hard to understand what it does. Start splitting it up, and pass arguments along -- as needed.
2017-03-30 12:23:26 +02:00
menu_switch_to_level0 (nmc->nmc_config.use_colors, &menu_ctx, BASE_PROMPT, nmc->editor_prompt_color);
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
nmc_tab_completion.setting = NULL; /* for TAB completion */
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
}
} else {
GError *tmp_err = NULL;
char *prop_name = is_property_valid (ss, cmd_arg_p, &tmp_err);
if (prop_name) {
/* Delete property value */
cli: deny removing values of nmcli-unchangable properties nmc_setting_reset_property() function checks whether we allow changing the property (set_func != NULL) and if so, the property value is reset to default.
2013-09-06 14:56:41 +02:00
if (!nmc_setting_reset_property (ss, prop_name, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: failed to remove value of '%s': %s\n"), prop_name,
tmp_err->message);
cli: deny removing values of nmcli-unchangable properties nmc_setting_reset_property() function checks whether we allow changing the property (set_func != NULL) and if so, the property value is reset to default.
2013-09-06 14:56:41 +02:00
g_clear_error (&tmp_err);
}
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
} else {
/* If the string is not a property, try it as a setting */
NMSetting *s_tmp;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
s_tmp = is_setting_valid (connection,
valid_settings_main,
valid_settings_slave,
cmd_arg_p);
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
if (s_tmp) {
/* Remove setting from the connection */
connection_remove_setting (connection, s_tmp);
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
/* coverity[copy_paste_error] - suppress Coverity COPY_PASTE_ERROR defect */
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
if (ss == menu_ctx.curr_setting) {
/* If we removed the setting we are in, go up */
cli: pass use_colors as separate option instead of global nmc nmc contains all options and collects output data. It is very hard to understand what it does. Start splitting it up, and pass arguments along -- as needed.
2017-03-30 12:23:26 +02:00
menu_switch_to_level0 (nmc->nmc_config.use_colors, &menu_ctx, BASE_PROMPT, nmc->editor_prompt_color);
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
nmc_tab_completion.setting = NULL; /* for TAB completion */
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
}
} else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: %s properties, nor it is a setting name.\n"),
tmp_err->message);
cli: add 'remove' command to the interactive editor remove <setting>[.property] | <property> It entirely removes given setting from edited connection. If a property is given, the command instead resets the property to its default value. https://bugzilla.gnome.org/show_bug.cgi?id=707576
2013-09-05 16:58:33 +02:00
g_clear_error (&tmp_err);
}
g_free (prop_name);
}
}
break;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
case NMC_EDITOR_MAIN_CMD_DESCRIBE:
/* Print property description */
if (!cmd_arg) {
if (menu_ctx.level == 1) {
const char *prop_name;
prop_name = ask_check_property (cmd_arg,
(const char **) menu_ctx.valid_props,
menu_ctx.valid_props_str);
if (!prop_name)
break;
/* Show property description */
print_property_description (menu_ctx.curr_setting, prop_name);
} else {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: no setting selected; valid are [%s]\n"), valid_settings_str);
g_print (_("use 'goto <setting>' first, or 'describe <setting>.<property>'\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
} else {
NMSetting *ss = NULL;
gboolean unref_ss = FALSE;
gboolean descr_all;
char *user_s;
/* cmd_arg_s != NULL means argument is "setting.property" */
descr_all = !cmd_arg_s && !menu_ctx.curr_setting;
user_s = descr_all ? cmd_arg_p : cmd_arg_s ? cmd_arg_s : NULL;
if (user_s) {
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
ss = is_setting_valid (connection,
valid_settings_main,
valid_settings_slave,
user_s);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (!ss) {
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
ss = create_setting_by_name (user_s,
valid_settings_main,
valid_settings_slave);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (!ss) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid setting argument '%s'; valid are [%s]\n"),
user_s, valid_settings_str);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
}
unref_ss = TRUE;
}
} else
ss = menu_ctx.curr_setting;
if (descr_all) {
/* Show description for all properties */
print_setting_description (ss);
} else {
GError *tmp_err = NULL;
char *prop_name = is_property_valid (ss, cmd_arg_p, &tmp_err);
if (prop_name) {
/* Show property description */
print_property_description (ss, prop_name);
} else {
/* If the string is not a property, try it as a setting */
NMSetting *s_tmp;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
s_tmp = is_setting_valid (connection,
valid_settings_main,
valid_settings_slave,
cmd_arg_p);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (s_tmp)
print_setting_description (s_tmp);
else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid property: %s, "
"neither a valid setting name.\n"),
tmp_err->message);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&tmp_err);
}
g_free (prop_name);
}
if (unref_ss)
g_object_unref (ss);
}
break;
case NMC_EDITOR_MAIN_CMD_PRINT:
/* Print current connection settings/properties */
if (cmd_arg) {
if (strcmp (cmd_arg, "all") == 0)
editor_show_connection (connection, nmc);
else {
cli: editor - enhance 'print' command - allow printing single properties: nmcli> print con.id connection.id: my-main-ethernet - allow printing other settings in second (settings) menu level: nmcli connection> print ipv4.method ipv4.method: auto nmcli connection> print eth ...
2014-06-18 13:41:47 +02:00
NMSetting *ss = NULL;
gboolean whole_setting;
char *user_s;
/* cmd_arg_s != NULL means argument is "setting.property" */
whole_setting = !cmd_arg_s && !menu_ctx.curr_setting;
user_s = whole_setting ? cmd_arg_p : cmd_arg_s ? cmd_arg_s : NULL;
if (user_s) {
const char *s_name;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
s_name = check_valid_name (user_s,
valid_settings_main,
valid_settings_slave,
NULL);
cli: editor - enhance 'print' command - allow printing single properties: nmcli> print con.id connection.id: my-main-ethernet - allow printing other settings in second (settings) menu level: nmcli connection> print ipv4.method ipv4.method: auto nmcli connection> print eth ...
2014-06-18 13:41:47 +02:00
if (!s_name) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: unknown setting: '%s'\n"), user_s);
cli: editor - enhance 'print' command - allow printing single properties: nmcli> print con.id connection.id: my-main-ethernet - allow printing other settings in second (settings) menu level: nmcli connection> print ipv4.method ipv4.method: auto nmcli connection> print eth ...
2014-06-18 13:41:47 +02:00
break;
}
ss = nm_connection_get_setting_by_name (connection, s_name);
if (!ss) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: '%s' setting not present in the connection\n"), s_name);
cli: editor - enhance 'print' command - allow printing single properties: nmcli> print con.id connection.id: my-main-ethernet - allow printing other settings in second (settings) menu level: nmcli connection> print ipv4.method ipv4.method: auto nmcli connection> print eth ...
2014-06-18 13:41:47 +02:00
break;
}
} else
ss = menu_ctx.curr_setting;
if (whole_setting) {
/* Print the whole setting */
editor_show_setting (ss, nmc);
} else {
GError *err = NULL;
char *prop_name = is_property_valid (ss, cmd_arg_p, &err);
if (prop_name) {
/* Print one property */
char *prop_val = nmc_setting_get_property (ss, prop_name, NULL);
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("%s.%s: %s\n", nm_setting_get_name (ss),prop_name , prop_val);
cli: editor - enhance 'print' command - allow printing single properties: nmcli> print con.id connection.id: my-main-ethernet - allow printing other settings in second (settings) menu level: nmcli connection> print ipv4.method ipv4.method: auto nmcli connection> print eth ...
2014-06-18 13:41:47 +02:00
g_free (prop_val);
} else {
/* If the string is not a property, try it as a setting */
NMSetting *s_tmp;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
s_tmp = is_setting_valid (connection,
valid_settings_main,
valid_settings_slave,
cmd_arg_p);
cli: editor - enhance 'print' command - allow printing single properties: nmcli> print con.id connection.id: my-main-ethernet - allow printing other settings in second (settings) menu level: nmcli connection> print ipv4.method ipv4.method: auto nmcli connection> print eth ...
2014-06-18 13:41:47 +02:00
if (s_tmp) {
/* Print the whole setting */
editor_show_setting (s_tmp, nmc);
} else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid property: %s%s\n"),
err->message,
cmd_arg_s ? "" : _(", neither a valid setting name"));
cli: editor - enhance 'print' command - allow printing single properties: nmcli> print con.id connection.id: my-main-ethernet - allow printing other settings in second (settings) menu level: nmcli connection> print ipv4.method ipv4.method: auto nmcli connection> print eth ...
2014-06-18 13:41:47 +02:00
g_clear_error (&err);
}
g_free (prop_name);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
}
} else {
if (menu_ctx.curr_setting)
editor_show_setting (menu_ctx.curr_setting, nmc);
else
editor_show_connection (connection, nmc);
}
break;
case NMC_EDITOR_MAIN_CMD_VERIFY:
/* Verify current setting or the whole connection */
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
if (cmd_arg && strcmp (cmd_arg, "all") && strcmp (cmd_arg, "fix")) {
g_print (_("Invalid verify option: %s\n"), cmd_arg);
break;
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if ( menu_ctx.curr_setting
&& (!cmd_arg || strcmp (cmd_arg, "all") != 0)) {
GError *tmp_err = NULL;
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
(void) nm_setting_verify (menu_ctx.curr_setting, NULL, &tmp_err);
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Verify setting '%s': %s\n"),
nm_setting_get_name (menu_ctx.curr_setting),
tmp_err ? tmp_err->message : "OK");
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&tmp_err);
} else {
GError *tmp_err = NULL;
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
gboolean valid, modified;
gboolean fixed = TRUE;
valid = nm_connection_verify (connection, &tmp_err);
if (!valid && (g_strcmp0 (cmd_arg, "fix") == 0)) {
/* Try to fix normalizable errors */
g_clear_error (&tmp_err);
fixed = nm_connection_normalize (connection, NULL, &modified, &tmp_err);
}
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Verify connection: %s\n"),
tmp_err ? tmp_err->message : "OK");
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
if (!fixed)
g_print (_("The error cannot be fixed automatically.\n"));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&tmp_err);
}
break;
case NMC_EDITOR_MAIN_CMD_SAVE:
/* Save the connection */
if (nm_connection_verify (connection, &err1)) {
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
gboolean persistent = TRUE;
/* parse argument */
if (cmd_arg) {
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (matches (cmd_arg, "temporary"))
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
persistent = FALSE;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
else if (matches (cmd_arg, "persistent"))
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
persistent = TRUE;
else {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid argument '%s'\n"), cmd_arg);
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
break;
}
}
cli: confirm connection saving when autoconnect=yes (rh #953296) Ask for confirmation in these cases: * autoconnect=yes and connection has not been saved yet * autoconnect=yes, but saved connection had autoconnect=no The confirmation can be switched off with nmcli> nmcli save-confirmation no
2013-07-23 14:25:52 +02:00
/* Ask for save confirmation if the connection changes to autoconnect=yes */
if (nmc->editor_save_confirmation)
if (!confirm_connection_saving (connection, NM_CONNECTION (rem_con)))
break;
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
if (!rem_con) {
/* Tell the settings service to add the new connection */
info = g_malloc0 (sizeof (AddConnectionInfo));
info->nmc = nmc;
info->con_name = g_strdup (nm_connection_get_id (connection));
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
add_new_connection (persistent,
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
nmc->client,
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
connection,
add_connection_editor_cb,
info);
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
} else {
/* Save/update already saved (existing) connection */
nm_connection_replace_settings_from_connection (NM_CONNECTION (rem_con),
libnm-core: fix up connection deserialize/copy/replace semantics libnm-util's connection deserializing/copying/replacing functions have odd semantics where sometimes they both modify a connection AND return an error. libnm-core tried to improve things by guaranteeing that the connection would not be modified if the new settings were invalid, but this ended up breaking a bunch of places that needed to be able to work with invalid connections. So re-fix the functions by reverting back to the old semantics, but having return values that clearly distinguish whether the connection was modified or not. For comparison: - nm_connection_new_from_hash() / nm_simple_connection_new_from_dbus(): - libnm-util: returns a valid connection or NULL. - OLD libnm-core: returned a valid connection or NULL. - NEW libnm-core: returns a valid connection or NULL. - nm_connection_duplicate() / nm_simple_connection_new_clone(): - libnm-util: always succeeds, whether or not the connection is valid. - OLD libnm-core: returned a valid connection or NULL - NEW libnm-core: always succeeds, whether or not the connection is valid. - nm_connection_replace_settings_from_connection(): - libnm-util: always replaces the settings, but returns FALSE if the connection is now invalid. - OLD libnm-core: either replaced the settings and returned TRUE (if the settings were valid), or else left the connection unchanged and returned FALSE (if not). - NEW libnm-core: always replaces the settings, and has no return value. (The modified connection is valid if and only if the replaced-from connection was valid; just like with the libnm-util version.) - nm_connection_replace_settings(): - libnm-util: returns TRUE if the new settings are valid, or FALSE if either (a) the new settings could not be deserialized and the connection is unchanged, or (b) the new settings were deserialized, and the connection was updated, but is now not valid. - OLD libnm-core: either replaced the settings and returned TRUE (if the settings were valid), or else left the connection unchanged and returned FALSE (if not). - NEW libnm-core: returns TRUE if the connection was updated (whether or not it is valid), or FALSE if the new settings could not be deserialized and the connection is unchanged.
2014-09-15 14:05:52 -04:00
connection);
cli: persistent vs. temporary save for 'nmcli con edit'
2014-01-24 14:48:30 +01:00
update_connection (persistent, rem_con, update_connection_editor_cb, NULL);
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
}
//FIXME: add also a timeout for cases the callback is not called
while (!nmc_editor_cb_called)
cli: remove editor thread Currently the editor runs in a dedicated thread so that the blocking call to readline() doesn't stop the processing of D-Bus events in the main loop. The editor thread can access objects concurrently with the main thread and this can cause races and crashes. Remove the editor thread and use the non-blocking readline API. https://bugzilla.gnome.org/show_bug.cgi?id=732097 https://bugzilla.redhat.com/show_bug.cgi?id=1368353
2016-09-02 16:16:34 +02:00
g_main_context_iteration (NULL, TRUE);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
if (nmc_editor_error) {
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
g_print (_("Error: Failed to save '%s' (%s) connection: %s\n"),
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
nm_connection_get_id (connection),
nm_connection_get_uuid (connection),
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
nmc_editor_error->message);
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
g_error_free (nmc_editor_error);
cli: copy remote connection to local one on 'save' (rh #997958) Plugins may have problems with preserving some properties on write/read cycle, may add ipv{4,6} settings when they are not present in the connection, etc. That makes local and remote connection differ. So we copy remote connection into the local to get rid of such problems. Note: 68f12b4e9c134d41e42cc8a1986cad6bb3c07b7a and f03635e5ac829d4fc896573f2f3c6969b9d5a2e2 commits ensure that all connection (except slaves) have IPv{4,6} settings. https://bugzilla.redhat.com/show_bug.cgi?id=997958
2013-10-08 12:24:05 +02:00
} else {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (!rem_con ?
_("Connection '%s' (%s) successfully saved.\n") :
_("Connection '%s' (%s) successfully updated.\n"),
nm_connection_get_id (connection),
nm_connection_get_uuid (connection));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
con_tmp = nm_client_get_connection_by_uuid (nmc->client,
nm_connection_get_uuid (connection));
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
g_weak_ref_set (&weak, con_tmp);
refresh_remote_connection (&weak, &rem_con);
cli: copy remote connection to local one on 'save' (rh #997958) Plugins may have problems with preserving some properties on write/read cycle, may add ipv{4,6} settings when they are not present in the connection, etc. That makes local and remote connection differ. So we copy remote connection into the local to get rid of such problems. Note: 68f12b4e9c134d41e42cc8a1986cad6bb3c07b7a and f03635e5ac829d4fc896573f2f3c6969b9d5a2e2 commits ensure that all connection (except slaves) have IPv{4,6} settings. https://bugzilla.redhat.com/show_bug.cgi?id=997958
2013-10-08 12:24:05 +02:00
/* Replace local connection with the remote one to be sure they are equal.
* This mitigates problems with plugins not preserving some properties or
* adding ipv{4,6} settings when not present.
*/
cli: fix TAB-completion and 'goto' after saving connection On 'save' we replace settings in the local connection, so we also need to update stored setting pointers.
2014-01-20 17:38:11 +01:00
if (con_tmp) {
char *s_name = NULL;
if (menu_ctx.curr_setting)
s_name = g_strdup (nm_setting_get_name (menu_ctx.curr_setting));
/* Update settings in the local connection */
cli: copy remote connection to local one on 'save' (rh #997958) Plugins may have problems with preserving some properties on write/read cycle, may add ipv{4,6} settings when they are not present in the connection, etc. That makes local and remote connection differ. So we copy remote connection into the local to get rid of such problems. Note: 68f12b4e9c134d41e42cc8a1986cad6bb3c07b7a and f03635e5ac829d4fc896573f2f3c6969b9d5a2e2 commits ensure that all connection (except slaves) have IPv{4,6} settings. https://bugzilla.redhat.com/show_bug.cgi?id=997958
2013-10-08 12:24:05 +02:00
nm_connection_replace_settings_from_connection (connection,
libnm-core: fix up connection deserialize/copy/replace semantics libnm-util's connection deserializing/copying/replacing functions have odd semantics where sometimes they both modify a connection AND return an error. libnm-core tried to improve things by guaranteeing that the connection would not be modified if the new settings were invalid, but this ended up breaking a bunch of places that needed to be able to work with invalid connections. So re-fix the functions by reverting back to the old semantics, but having return values that clearly distinguish whether the connection was modified or not. For comparison: - nm_connection_new_from_hash() / nm_simple_connection_new_from_dbus(): - libnm-util: returns a valid connection or NULL. - OLD libnm-core: returned a valid connection or NULL. - NEW libnm-core: returns a valid connection or NULL. - nm_connection_duplicate() / nm_simple_connection_new_clone(): - libnm-util: always succeeds, whether or not the connection is valid. - OLD libnm-core: returned a valid connection or NULL - NEW libnm-core: always succeeds, whether or not the connection is valid. - nm_connection_replace_settings_from_connection(): - libnm-util: always replaces the settings, but returns FALSE if the connection is now invalid. - OLD libnm-core: either replaced the settings and returned TRUE (if the settings were valid), or else left the connection unchanged and returned FALSE (if not). - NEW libnm-core: always replaces the settings, and has no return value. (The modified connection is valid if and only if the replaced-from connection was valid; just like with the libnm-util version.) - nm_connection_replace_settings(): - libnm-util: returns TRUE if the new settings are valid, or FALSE if either (a) the new settings could not be deserialized and the connection is unchanged, or (b) the new settings were deserialized, and the connection was updated, but is now not valid. - OLD libnm-core: either replaced the settings and returned TRUE (if the settings were valid), or else left the connection unchanged and returned FALSE (if not). - NEW libnm-core: returns TRUE if the connection was updated (whether or not it is valid), or FALSE if the new settings could not be deserialized and the connection is unchanged.
2014-09-15 14:05:52 -04:00
NM_CONNECTION (con_tmp));
cli: fix TAB-completion and 'goto' after saving connection On 'save' we replace settings in the local connection, so we also need to update stored setting pointers.
2014-01-20 17:38:11 +01:00
/* Also update setting for menu context and TAB-completion */
menu_ctx.curr_setting = s_name ? nm_connection_get_setting_by_name (connection, s_name) : NULL;
nmc_tab_completion.setting = menu_ctx.curr_setting;
g_free (s_name);
}
cli: copy remote connection to local one on 'save' (rh #997958) Plugins may have problems with preserving some properties on write/read cycle, may add ipv{4,6} settings when they are not present in the connection, etc. That makes local and remote connection differ. So we copy remote connection into the local to get rid of such problems. Note: 68f12b4e9c134d41e42cc8a1986cad6bb3c07b7a and f03635e5ac829d4fc896573f2f3c6969b9d5a2e2 commits ensure that all connection (except slaves) have IPv{4,6} settings. https://bugzilla.redhat.com/show_bug.cgi?id=997958
2013-10-08 12:24:05 +02:00
}
cli: fix saving connections in the connection editor - Main command loop was moved to run in a separate thread (using GThread). Otherwise glib main loop would be blocked in processing D-Bus and other events. - Handle creating new vs. updating already saved connection.
2013-03-21 17:08:44 +01:00
nmc_editor_cb_called = FALSE;
nmc_editor_error = NULL;
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
} else {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: connection verification failed: %s\n"),
err1 ? err1->message : _("(unknown error)"));
cli: add 'verify fix' command for nmcli connection editor Some connection verification errors are normalizable by NetworkManager. So we can allow users to normalize the connection using one command.
2014-10-27 11:03:19 +01:00
g_print (_("You may try running 'verify fix' to fix errors.\n"));
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_clear_error (&err1);
break;
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
case NMC_EDITOR_MAIN_CMD_ACTIVATE:
{
GError *tmp_err = NULL;
const char *ifname = cmd_arg_p;
const char *ap_nsp = cmd_arg_v;
/* When only AP/NSP is specified it is prepended with '/' */
if (!cmd_arg_v) {
if (ifname && ifname[0] == '/') {
ap_nsp = ifname + 1;
ifname = NULL;
}
} else
ap_nsp = ap_nsp && ap_nsp[0] == '/' ? ap_nsp + 1 : ap_nsp;
cli: always get up-to-date dirty status
2014-01-20 18:06:37 +01:00
if (is_connection_dirty (connection, rem_con)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: connection is not saved. Type 'save' first.\n"));
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
break;
}
if (!nm_connection_verify (NM_CONNECTION (rem_con), &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: connection is not valid: %s\n"), tmp_err->message);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
g_clear_error (&tmp_err);
break;
}
nmc->nowait_flag = FALSE;
cli: turn should_wait into a semaphore Count the exit blockers. We'll want to terminate the device monitor in case all monitored devices vanish.
2015-03-27 12:39:17 +01:00
nmc->should_wait++;
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
nmc->nmc_config_mutable.print_output = NMC_PRINT_PRETTY;
cli: add 'passwd-file' option for 'nmcli connection up' to provide passwords It is useful for running nmcli without --ask option, i.e. non-interactively. Example contents of the file: wifi.psk: s e c r e t 12345 802-1x.password:kili manjaro 802-1x.pin:987654321
2014-10-14 15:53:34 +02:00
if (!nmc_activate_connection (nmc, NM_CONNECTION (rem_con), ifname, ap_nsp, ap_nsp, NULL,
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
activate_connection_editor_cb, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: Cannot activate connection: %s.\n"), tmp_err->message);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
g_clear_error (&tmp_err);
break;
}
while (!nmc_editor_cb_called)
cli: remove editor thread Currently the editor runs in a dedicated thread so that the blocking call to readline() doesn't stop the processing of D-Bus events in the main loop. The editor thread can access objects concurrently with the main thread and this can cause races and crashes. Remove the editor thread and use the non-blocking readline API. https://bugzilla.gnome.org/show_bug.cgi?id=732097 https://bugzilla.redhat.com/show_bug.cgi?id=1368353
2016-09-02 16:16:34 +02:00
g_main_context_iteration (NULL, TRUE);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
if (nmc_editor_error) {
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
g_print (_("Error: Failed to activate '%s' (%s) connection: %s\n"),
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
nm_connection_get_id (connection),
nm_connection_get_uuid (connection),
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
nmc_editor_error->message);
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
g_error_free (nmc_editor_error);
} else {
cli: remove editor thread Currently the editor runs in a dedicated thread so that the blocking call to readline() doesn't stop the processing of D-Bus events in the main loop. The editor thread can access objects concurrently with the main thread and this can cause races and crashes. Remove the editor thread and use the non-blocking readline API. https://bugzilla.gnome.org/show_bug.cgi?id=732097 https://bugzilla.redhat.com/show_bug.cgi?id=1368353
2016-09-02 16:16:34 +02:00
nmc_readline (_("Monitoring connection activation (press any key to continue)\n"));
cli: add 'activate' menu command for interactive editor (rh #1004883) This command allows activating the edited connection. Monitoring the progress of the activation is a bit complicated by the fact that the callback activate_connection_editor_cb() is invoked in main loop thread, and not in the editor thread itself. https://bugzilla.redhat.com/show_bug.cgi?id=1004883
2013-07-23 13:32:10 +02:00
}
if (nmc_editor_monitor_ac) {
if (nmc_editor_monitor_ac->monitor_id)
g_source_remove (nmc_editor_monitor_ac->monitor_id);
g_free (nmc_editor_monitor_ac);
}
nmc_editor_cb_called = FALSE;
nmc_editor_error = NULL;
nmc_editor_monitor_ac = NULL;
/* Update timestamp in local connection */
update_connection_timestamp (NM_CONNECTION (rem_con), connection);
}
break;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
case NMC_EDITOR_MAIN_CMD_BACK:
/* Go back (up) an the menu */
if (menu_ctx.level == 1) {
cli: pass use_colors as separate option instead of global nmc nmc contains all options and collects output data. It is very hard to understand what it does. Start splitting it up, and pass arguments along -- as needed.
2017-03-30 12:23:26 +02:00
menu_switch_to_level0 (nmc->nmc_config.use_colors, &menu_ctx, BASE_PROMPT, nmc->editor_prompt_color);
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
nmc_tab_completion.setting = NULL; /* for TAB completion */
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
break;
case NMC_EDITOR_MAIN_CMD_HELP:
/* Print command help */
editor_main_help (cmd_arg);
break;
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
case NMC_EDITOR_MAIN_CMD_NMCLI:
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (cmd_arg_p && matches (cmd_arg_p, "status-line")) {
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
GError *tmp_err = NULL;
gboolean bb;
if (!nmc_string_to_bool (cmd_arg_v ? g_strstrip (cmd_arg_v) : "", &bb, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: status-line: %s\n"), tmp_err->message);
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
g_clear_error (&tmp_err);
} else
nmc->editor_status_line = bb;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
} else if (cmd_arg_p && matches (cmd_arg_p, "save-confirmation")) {
cli: confirm connection saving when autoconnect=yes (rh #953296) Ask for confirmation in these cases: * autoconnect=yes and connection has not been saved yet * autoconnect=yes, but saved connection had autoconnect=no The confirmation can be switched off with nmcli> nmcli save-confirmation no
2013-07-23 14:25:52 +02:00
GError *tmp_err = NULL;
gboolean bb;
if (!nmc_string_to_bool (cmd_arg_v ? g_strstrip (cmd_arg_v) : "", &bb, &tmp_err)) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: save-confirmation: %s\n"), tmp_err->message);
cli: confirm connection saving when autoconnect=yes (rh #953296) Ask for confirmation in these cases: * autoconnect=yes and connection has not been saved yet * autoconnect=yes, but saved connection had autoconnect=no The confirmation can be switched off with nmcli> nmcli save-confirmation no
2013-07-23 14:25:52 +02:00
g_clear_error (&tmp_err);
} else
nmc->editor_save_confirmation = bb;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
} else if (cmd_arg_p && matches (cmd_arg_p, "show-secrets")) {
cli: allow showing secrets in the editor
2014-09-29 13:57:22 +02:00
GError *tmp_err = NULL;
gboolean bb;
if (!nmc_string_to_bool (cmd_arg_v ? g_strstrip (cmd_arg_v) : "", &bb, &tmp_err)) {
g_print (_("Error: show-secrets: %s\n"), tmp_err->message);
g_clear_error (&tmp_err);
} else
nmc->editor_show_secrets = bb;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
} else if (cmd_arg_p && matches (cmd_arg_p, "prompt-color")) {
cli: take color name arguments for "nmcli prompt-color" in editor (bgo #744936) Adjust nmcli prompt-color description and make it more friendly for translators. nmcli> nmcli prompt-color green https://bugzilla.gnome.org/show_bug.cgi?id=744936
2015-06-10 17:41:53 +02:00
GError *tmp_err = NULL;
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
NMMetaTermColor color;
cli: take color name arguments for "nmcli prompt-color" in editor (bgo #744936) Adjust nmcli prompt-color description and make it more friendly for translators. nmcli> nmcli prompt-color green https://bugzilla.gnome.org/show_bug.cgi?id=744936
2015-06-10 17:41:53 +02:00
color = nmc_term_color_parse_string (cmd_arg_v ? g_strstrip (cmd_arg_v) : " ", &tmp_err);
if (tmp_err) {
g_print (_("Error: bad color: %s\n"), tmp_err->message);
g_clear_error (&tmp_err);
} else {
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
nmc->editor_prompt_color = color;
g_free (menu_ctx.main_prompt);
if (menu_ctx.level == 0)
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
menu_ctx.main_prompt = nmc_colorize (nmc->nmc_config.use_colors, nmc->editor_prompt_color, NM_META_TERM_FORMAT_NORMAL,
nmcli: add support for terminal formatting, like bold, dim, underline, etc.
2015-02-17 13:20:37 +01:00
BASE_PROMPT);
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
else
cli: move and rename TermColor and TermFormat
2017-04-04 15:12:00 +02:00
menu_ctx.main_prompt = nmc_colorize (nmc->nmc_config.use_colors, nmc->editor_prompt_color, NM_META_TERM_FORMAT_NORMAL,
nmcli: add support for terminal formatting, like bold, dim, underline, etc.
2015-02-17 13:20:37 +01:00
"nmcli %s> ",
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
nm_setting_get_name (menu_ctx.curr_setting));
}
cli: editor: show current configuration for 'nmcli' command without arguments
2013-07-23 15:29:48 +02:00
} else if (!cmd_arg_p) {
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Current nmcli configuration:\n"));
g_print ("status-line: %s\n"
"save-confirmation: %s\n"
cli: allow showing secrets in the editor
2014-09-29 13:57:22 +02:00
"show-secrets: %s\n"
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
"prompt-color: %d\n",
nmc->editor_status_line ? "yes" : "no",
nmc->editor_save_confirmation ? "yes" : "no",
cli: allow showing secrets in the editor
2014-09-29 13:57:22 +02:00
nmc->editor_show_secrets ? "yes" : "no",
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
nmc->editor_prompt_color);
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
} else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Invalid configuration option '%s'; allowed [%s]\n"),
cli: allow showing secrets in the editor
2014-09-29 13:57:22 +02:00
cmd_arg_v ? cmd_arg_v : "", "status-line, save-confirmation, show-secrets, prompt-color");
cli: add 'nmcli' command for nmcli connection editor configuration - show/hide status line: nmcli> nmcli status-line yes | no - editor prompt color: nmcli> nmcli prompt-color <0-8> (0 = normal/without color)
2013-04-05 23:42:56 +02:00
break;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
case NMC_EDITOR_MAIN_CMD_QUIT:
cli: always get up-to-date dirty status
2014-01-20 18:06:37 +01:00
if (is_connection_dirty (connection, rem_con)) {
cli: use readline for quit confirmation in editor, and accept localized answer
2014-06-02 12:24:56 +02:00
if (confirm_quit ())
cli: use 'dirty' flag for "Do you really want to quit?" confirmation in editor
2013-04-23 12:53:52 +02:00
cmd_loop = FALSE; /* quit command loop */
} else
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cmd_loop = FALSE; /* quit command loop */
break;
case NMC_EDITOR_MAIN_CMD_UNKNOWN:
default:
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Unknown command: '%s'\n"), cmd_user);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
break;
}
g_free (cmd_user);
g_free (cmd_arg);
g_free (cmd_arg_s);
g_free (cmd_arg_p);
g_free (cmd_arg_v);
}
g_free (valid_settings_str);
g_free (menu_ctx.main_prompt);
g_strfreev (menu_ctx.valid_props);
g_free (menu_ctx.valid_props_str);
cli: fix crash when edited profile was removed by another client (rh #1011942) Use g_weak_ref_get() that either returns an object with reference increment or returns NULL. That fixes the problem. However, in the long run we should rework the editor loop trying to merge that with GMainLoop, which could help for various issues. https://bugzilla.redhat.com/show_bug.cgi?id=1011942
2014-01-17 12:00:03 +01:00
if (rem_con)
g_object_unref (rem_con);
g_weak_ref_clear (&weak);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: load/save nmcli editor (readline) history so that it is persistent The history is saved to ~/.nmcli-history file, separately for each connection. The file uses glib key-file format. Each group is a connection UUID, keys are integer numbers (history entry order), and values are the actual commands. Example ~/.nmcli-history file: [0bdc9852-2540-4e12-a605-5e65a9483772] 0=help quit 1=print 2=nmcli prompt-color 3 3=help set 4=q [9142680d-3b87-4feb-ab1e-19e8762329ad] 0=eth 1=set ipv4.addr 1.2.3.4 2=quit
2013-06-24 10:29:49 +02:00
/* Save history file */
save_history_cmds (nm_connection_get_uuid (connection));
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
return TRUE;
}
static const char *
get_ethernet_device_name (NmCli *nmc)
{
const GPtrArray *devices;
int i;
devices = nm_client_get_devices (nmc->client);
libnm: change empty-GPtrArray-return semantics libnm functions that return GPtrArrays of objects had a rule that if the array was empty, they would return NULL rather than a 0-length array. As it turns out, this is just a nuisance to clients, since in most places the code for the non-empty case would end up doing the right thing for the empty case as well (and where it doesn't, we can check "array->len == 0" just as easily as "array == NULL"). So just return the 0-length array instead.
2014-08-25 10:33:53 -04:00
for (i = 0; i < devices->len; i++) {
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NMDevice *dev = g_ptr_array_index (devices, i);
if (NM_IS_DEVICE_ETHERNET (dev))
return nm_device_get_iface (dev);
}
return NULL;
}
static void
editor_init_new_connection (NmCli *nmc, NMConnection *connection)
{
cli: fix new connection initialization in the editor This fixes a regression caused by ecd49fb435cd15aafb9a46c85ed5c42c525d0c15 In addition, initialize Wi-Fi mode to "infrastructure".
2013-10-25 14:18:15 +02:00
NMSetting *setting, *base_setting;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NMSettingConnection *s_con;
const char *con_type;
const char *slave_type = NULL;
s_con = nm_connection_get_setting_connection (connection);
g_assert (s_con);
con_type = nm_setting_connection_get_connection_type (s_con);
cli: properly initialize new team-slave connections in editor
2013-10-21 16:27:23 +02:00
/* Initialize new connection according to its type using sensible defaults. */
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: check IP configuration when setting connection.master in editor If IPv4 and/or IPv6 are present on setting connection.master, offer their removal.
2014-10-27 12:55:55 +01:00
nmc_setting_connection_connect_handlers (s_con, connection);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (g_strcmp0 (con_type, "bond-slave") == 0)
slave_type = NM_SETTING_BOND_SETTING_NAME;
cli: properly initialize new team-slave connections in editor
2013-10-21 16:27:23 +02:00
if (g_strcmp0 (con_type, "team-slave") == 0)
slave_type = NM_SETTING_TEAM_SETTING_NAME;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
if (g_strcmp0 (con_type, "bridge-slave") == 0)
slave_type = NM_SETTING_BRIDGE_SETTING_NAME;
if (slave_type) {
const char *dev_ifname = get_ethernet_device_name (nmc);
cli: properly initialize new team-slave connections in editor
2013-10-21 16:27:23 +02:00
/* For bond/team/bridge slaves add 'wired' setting */
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
setting = nm_setting_wired_new ();
nm_connection_add_setting (connection, setting);
g_object_set (s_con,
NM_SETTING_CONNECTION_TYPE, NM_SETTING_WIRED_SETTING_NAME,
NM_SETTING_CONNECTION_MASTER, dev_ifname ? dev_ifname : "eth0",
NM_SETTING_CONNECTION_SLAVE_TYPE, slave_type,
NULL);
cli: properly initialize new team-slave connections in editor
2013-10-21 16:27:23 +02:00
} else {
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* Add a "base" setting to the connection by default */
cli: fix new connection initialization in the editor This fixes a regression caused by ecd49fb435cd15aafb9a46c85ed5c42c525d0c15 In addition, initialize Wi-Fi mode to "infrastructure".
2013-10-25 14:18:15 +02:00
base_setting = nmc_setting_new_for_name (con_type);
if (!base_setting)
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
return;
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
nmc_setting_custom_init (base_setting);
cli: fix new connection initialization in the editor This fixes a regression caused by ecd49fb435cd15aafb9a46c85ed5c42c525d0c15 In addition, initialize Wi-Fi mode to "infrastructure".
2013-10-25 14:18:15 +02:00
nm_connection_add_setting (connection, base_setting);
cli: add IPv{4,6} setting for all new non-slave connections in editor
2013-10-21 17:02:27 +02:00
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
set_default_interface_name (nmc, s_con);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* Set sensible initial VLAN values */
if (g_strcmp0 (con_type, NM_SETTING_VLAN_SETTING_NAME) == 0) {
const char *dev_ifname = get_ethernet_device_name (nmc);
cli: fix new connection initialization in the editor This fixes a regression caused by ecd49fb435cd15aafb9a46c85ed5c42c525d0c15 In addition, initialize Wi-Fi mode to "infrastructure".
2013-10-25 14:18:15 +02:00
g_object_set (NM_SETTING_VLAN (base_setting),
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NM_SETTING_VLAN_PARENT, dev_ifname ? dev_ifname : "eth0",
cli: fix new connection initialization in the editor This fixes a regression caused by ecd49fb435cd15aafb9a46c85ed5c42c525d0c15 In addition, initialize Wi-Fi mode to "infrastructure".
2013-10-25 14:18:15 +02:00
NULL);
cli: show warning when setting band/channel for infra mode (rh #1000096) The band/channel property is not considered when connecting to an access point. Print a warning, when setting one of these values in edit mode. For now, don't forbid the user to make such setting. Also, because in the future, wpa_supplicant might support this. https://bugzilla.redhat.com/show_bug.cgi?id=999997 https://bugzilla.redhat.com/show_bug.cgi?id=1000096 Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-02-13 12:20:52 +01:00
}
cli: initialize adsl.protocol in editor when creating a new connection The protocol is required and it is nice to have a valid initial value for the property.
2015-09-18 11:08:28 +02:00
cli: fix new connection initialization in the editor This fixes a regression caused by ecd49fb435cd15aafb9a46c85ed5c42c525d0c15 In addition, initialize Wi-Fi mode to "infrastructure".
2013-10-25 14:18:15 +02:00
/* Always add IPv4 and IPv6 settings for non-slave connections */
setting = nm_setting_ip4_config_new ();
nmc_setting_custom_init (setting);
nm_connection_add_setting (connection, setting);
setting = nm_setting_ip6_config_new ();
nmc_setting_custom_init (setting);
nm_connection_add_setting (connection, setting);
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
/* Also Proxy Setting */
setting = nm_setting_proxy_new ();
nmc_setting_custom_init (setting);
nm_connection_add_setting (connection, setting);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
}
cli: connect IP addresses/method handlers for existing connections (rh #998137) The handlers detecting changes of IP addresses/method in nmcli interactive editor were connected only for newly created connection. That's why the automagic feature of setting 'method' when 'addresses' are changed and vice versa worked just for new connections, but not while editing existing connections.
2013-09-02 13:20:04 +02:00
static void
editor_init_existing_connection (NMConnection *connection)
{
libnm-core: extract NMSettingIPConfig superclass out of IP4, IP6 classes Split a base NMSettingIPConfig class out of NMSettingIP4Config and NMSettingIP6Config, and update things accordingly. Further simplifications of now-redundant IPv4-vs-IPv6 code are possible, and should happen in the future.
2014-10-19 17:30:10 -04:00
NMSettingIPConfig *s_ip4, *s_ip6;
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
NMSettingProxy *s_proxy;
cli: show warning when setting band/channel for infra mode (rh #1000096) The band/channel property is not considered when connecting to an access point. Print a warning, when setting one of these values in edit mode. For now, don't forbid the user to make such setting. Also, because in the future, wpa_supplicant might support this. https://bugzilla.redhat.com/show_bug.cgi?id=999997 https://bugzilla.redhat.com/show_bug.cgi?id=1000096 Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-02-13 12:20:52 +01:00
NMSettingWireless *s_wireless;
cli: check IP configuration when setting connection.master in editor If IPv4 and/or IPv6 are present on setting connection.master, offer their removal.
2014-10-27 12:55:55 +01:00
NMSettingConnection *s_con;
cli: connect IP addresses/method handlers for existing connections (rh #998137) The handlers detecting changes of IP addresses/method in nmcli interactive editor were connected only for newly created connection. That's why the automagic feature of setting 'method' when 'addresses' are changed and vice versa worked just for new connections, but not while editing existing connections.
2013-09-02 13:20:04 +02:00
s_ip4 = nm_connection_get_setting_ip4_config (connection);
s_ip6 = nm_connection_get_setting_ip6_config (connection);
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
s_proxy = nm_connection_get_setting_proxy (connection);
cli: show warning when setting band/channel for infra mode (rh #1000096) The band/channel property is not considered when connecting to an access point. Print a warning, when setting one of these values in edit mode. For now, don't forbid the user to make such setting. Also, because in the future, wpa_supplicant might support this. https://bugzilla.redhat.com/show_bug.cgi?id=999997 https://bugzilla.redhat.com/show_bug.cgi?id=1000096 Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-02-13 12:20:52 +01:00
s_wireless = nm_connection_get_setting_wireless (connection);
cli: check IP configuration when setting connection.master in editor If IPv4 and/or IPv6 are present on setting connection.master, offer their removal.
2014-10-27 12:55:55 +01:00
s_con = nm_connection_get_setting_connection (connection);
cli: connect IP addresses/method handlers for existing connections (rh #998137) The handlers detecting changes of IP addresses/method in nmcli interactive editor were connected only for newly created connection. That's why the automagic feature of setting 'method' when 'addresses' are changed and vice versa worked just for new connections, but not while editing existing connections.
2013-09-02 13:20:04 +02:00
if (s_ip4)
nmc_setting_ip4_connect_handlers (s_ip4);
if (s_ip6)
nmc_setting_ip6_connect_handlers (s_ip6);
nmcli: Add Proxy support to nmcli nmcli fixed to allow users edit proxy properties.
2016-08-16 05:56:00 +05:30
if (s_proxy)
nmc_setting_proxy_connect_handlers (s_proxy);
cli: show warning when setting band/channel for infra mode (rh #1000096) The band/channel property is not considered when connecting to an access point. Print a warning, when setting one of these values in edit mode. For now, don't forbid the user to make such setting. Also, because in the future, wpa_supplicant might support this. https://bugzilla.redhat.com/show_bug.cgi?id=999997 https://bugzilla.redhat.com/show_bug.cgi?id=1000096 Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-02-13 12:20:52 +01:00
if (s_wireless)
nmc_setting_wireless_connect_handlers (s_wireless);
cli: check IP configuration when setting connection.master in editor If IPv4 and/or IPv6 are present on setting connection.master, offer their removal.
2014-10-27 12:55:55 +01:00
if (s_con)
nmc_setting_connection_connect_handlers (s_con, connection);
cli: connect IP addresses/method handlers for existing connections (rh #998137) The handlers detecting changes of IP addresses/method in nmcli interactive editor were connected only for newly created connection. That's why the automagic feature of setting 'method' when 'addresses' are changed and vice versa worked just for new connections, but not while editing existing connections.
2013-09-02 13:20:04 +02:00
}
cli: add completion for "nmcli c edit"
2017-02-14 15:57:08 +01:00
static void
nmc_complete_connection_type (const char *prefix, const NameItem *types)
{
while (types->name) {
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (!*prefix || matches (prefix, types->name))
cli: add completion for "nmcli c edit"
2017-02-14 15:57:08 +01:00
g_print ("%s\n", types->name);
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (types->alias && (!*prefix || matches (prefix, types->alias)))
cli: add completion for "nmcli c edit"
2017-02-14 15:57:08 +01:00
g_print ("%s\n", types->alias);
types++;
}
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
static NMCResultCode
cli: remove editor thread Currently the editor runs in a dedicated thread so that the blocking call to readline() doesn't stop the processing of D-Bus events in the main loop. The editor thread can access objects concurrently with the main thread and this can cause races and crashes. Remove the editor thread and use the non-blocking readline API. https://bugzilla.gnome.org/show_bug.cgi?id=732097 https://bugzilla.redhat.com/show_bug.cgi?id=1368353
2016-09-02 16:16:34 +02:00
do_connection_edit (NmCli *nmc, int argc, char **argv)
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
NMConnection *connection = NULL;
NMSettingConnection *s_con;
const char *connection_type;
char *uuid;
char *default_name = NULL;
const char *type = NULL;
char *type_ask = NULL;
const char *con_name = NULL;
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
const char *con = NULL;
const char *con_id = NULL;
const char *con_uuid = NULL;
const char *con_path = NULL;
const char *selector = NULL;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
char *tmp_str;
GError *error = NULL;
GError *err1 = NULL;
nmc_arg_t exp_args[] = { {"type", TRUE, &type, FALSE},
{"con-name", TRUE, &con_name, FALSE},
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
{"id", TRUE, &con_id, FALSE},
{"uuid", TRUE, &con_uuid, FALSE},
{"path", TRUE, &con_path, FALSE},
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
{NULL} };
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli: add completion for "nmcli c edit"
2017-02-14 15:57:08 +01:00
if (argc == 1 && nmc->complete)
nmc_complete_strings (*argv, "type", "con-name", "id", "uuid", "path", NULL);
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
nmc->return_value = NMC_RESULT_SUCCESS;
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
if (argc == 1)
con = *argv;
else {
if (!nmc_parse_args (exp_args, TRUE, &argc, &argv, &error)) {
g_string_assign (nmc->return_text, error->message);
nmc->return_value = error->code;
g_clear_error (&error);
goto error;
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
}
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
/* Setup some readline completion stuff */
/* Set a pointer to an alternative function to create matches */
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
rl_attempted_completion_function = (rl_completion_func_t *) nmcli_editor_tab_completion;
cli/build: make libreadline a build-time dependency (bgo #729846) and link nmcli against it. This simplifies code, makes sure that line editing capability is supported, and will allow using readline on other places too. https://bugzilla.gnome.org/show_bug.cgi?id=729846 https://bugzilla.redhat.com/show_bug.cgi?id=1007365
2014-05-13 15:50:38 +02:00
/* Use ' ' and '.' as word break characters */
rl_completer_word_break_characters = ". ";
cli: line-editing capability for interactive editor using an edit line library Supported libraries are: GNU Readline (libreadline) http://cnswww.cns.cwru.edu/php/chet/readline/rltop.html NetBSD Editline (libedit) http://www.thrysoee.dk/editline/ We load these symbols: readline() - main function allowing getting user input and aditing that add_history() - stores strings to history rl_insert_text() - insert text into the line at the current cursor position rl_startup_hook - hook to call just before readline prints the first prompt
2013-01-31 19:03:51 +01:00
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
if (!con) {
if (con_id && !con_uuid && !con_path) {
con = con_id;
selector = "id";
} else if (con_uuid && !con_id && !con_path) {
con = con_uuid;
selector = "uuid";
} else if (con_path && !con_id && !con_uuid) {
con = con_path;
selector = "path";
} else if (!con_path && !con_id && !con_uuid) {
/* no-op */
} else {
g_string_printf (nmc->return_text,
_("Error: only one of 'id', uuid, or 'path' can be provided."));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto error;
}
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
if (con) {
/* Existing connection */
NMConnection *found_con;
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: add completion for "nmcli c edit"
2017-02-14 15:57:08 +01:00
found_con = nmc_find_connection (connections, selector, con, NULL, nmc->complete);
if (nmc->complete)
goto error;
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
if (!found_con) {
g_string_printf (nmc->return_text, _("Error: Unknown connection '%s'."), con);
cli: add NMC_RESULT_ERROR_NOT_FOUND and use as appropriate This will allow callers to distinguish "no such connection" from "connection failed to activate", etc. https://bugzilla.redhat.com/show_bug.cgi?id=1022256
2013-11-21 09:19:18 -05:00
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
goto error;
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
/* Duplicate the connection and use that so that we need not
* differentiate existing vs. new later
*/
libnm-core, libnm, core: make NMConnection an interface The fact that NMRemoteConnection has to be an NMConnection and therefore can't be an NMObject means that it needs to reimplement bits of NMObject functionality (and likewise NMObject needs some special magic to deal with it). Likewise, we will need a daemon-side equivalent of NMObject as part of the gdbus port, and we would want NMSettingsConnection to be able to inherit from this as well. Solve this problem by making NMConnection into an interface, and having NMRemoteConnection and NMSettingsConnection implement it. (We use some hacks to keep the GHashTable of NMSettings objects inside nm-connection.c rather than having to be implemented by the implementations.) Since NMConnection is no longer an instantiable type, this adds NMSimpleConnection to replace the various non-D-Bus-based uses of NMConnection throughout the code. nm_connection_new() becomes nm_simple_connection_new(), nm_connection_new_from_hash() becomes nm_simple_connection_new_from_hash(), and nm_connection_duplicate() becomes nm_simple_connection_new_clone().
2014-08-13 14:34:29 -04:00
connection = nm_simple_connection_new_clone (found_con);
cli: fix showing secrets in nmcli editor (bgo #737415) Put secrets into the local connection. https://bugzilla.gnome.org/show_bug.cgi?id=737415
2014-11-06 17:16:01 +01:00
/* Merge secrets into the connection */
update_secrets_in_connection (NM_REMOTE_CONNECTION (found_con), connection);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
s_con = nm_connection_get_setting_connection (connection);
g_assert (s_con);
connection_type = nm_setting_connection_get_connection_type (s_con);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
if (type)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Warning: editing existing connection '%s'; 'type' argument is ignored\n"),
nm_connection_get_id (connection));
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
if (con_name)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Warning: editing existing connection '%s'; 'con-name' argument is ignored\n"),
nm_connection_get_id (connection));
cli: load/save nmcli editor (readline) history so that it is persistent The history is saved to ~/.nmcli-history file, separately for each connection. The file uses glib key-file format. Each group is a connection UUID, keys are integer numbers (history entry order), and values are the actual commands. Example ~/.nmcli-history file: [0bdc9852-2540-4e12-a605-5e65a9483772] 0=help quit 1=print 2=nmcli prompt-color 3 3=help set 4=q [9142680d-3b87-4feb-ab1e-19e8762329ad] 0=eth 1=set ipv4.addr 1.2.3.4 2=quit
2013-06-24 10:29:49 +02:00
/* Load previously saved history commands for the connection */
load_history_cmds (nm_connection_get_uuid (connection));
cli: connect IP addresses/method handlers for existing connections (rh #998137) The handlers detecting changes of IP addresses/method in nmcli interactive editor were connected only for newly created connection. That's why the automagic feature of setting 'method' when 'addresses' are changed and vice versa worked just for new connections, but not while editing existing connections.
2013-09-02 13:20:04 +02:00
editor_init_existing_connection (connection);
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
} else {
/* New connection */
cli: add completion for "nmcli c edit"
2017-02-14 15:57:08 +01:00
if (nmc->complete) {
if (type && argc == 0)
nmc_complete_connection_type (type, nmc_valid_connection_types);
goto error;
}
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
connection_type = check_valid_name (type, nmc_valid_connection_types, NULL, &err1);
tmp_str = get_valid_options_string (nmc_valid_connection_types, NULL);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
while (!connection_type) {
if (!type)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Valid connection types: %s\n"), tmp_str);
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Error: invalid connection type; %s\n"), err1->message);
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
g_clear_error (&err1);
cli: make prompt argument a printf-like format string, remove history argument for nmc_readline()
2014-06-02 10:18:13 +02:00
type_ask = nmc_readline (EDITOR_PROMPT_CON_TYPE);
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
type = type_ask = type_ask ? g_strstrip (type_ask) : NULL;
nmcli: fix check on allowed parameters for slave connections When editing a connection, check its slave-type parameter. If not null, we will not have ip properties but possibly others, slave-type specific.
2016-02-23 11:56:46 +01:00
connection_type = check_valid_name (type_ask, nmc_valid_connection_types, NULL, &err1);
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
g_free (type_ask);
}
g_free (tmp_str);
/* Create a new connection object */
libnm-core, libnm, core: make NMConnection an interface The fact that NMRemoteConnection has to be an NMConnection and therefore can't be an NMObject means that it needs to reimplement bits of NMObject functionality (and likewise NMObject needs some special magic to deal with it). Likewise, we will need a daemon-side equivalent of NMObject as part of the gdbus port, and we would want NMSettingsConnection to be able to inherit from this as well. Solve this problem by making NMConnection into an interface, and having NMRemoteConnection and NMSettingsConnection implement it. (We use some hacks to keep the GHashTable of NMSettings objects inside nm-connection.c rather than having to be implemented by the implementations.) Since NMConnection is no longer an instantiable type, this adds NMSimpleConnection to replace the various non-D-Bus-based uses of NMConnection throughout the code. nm_connection_new() becomes nm_simple_connection_new(), nm_connection_new_from_hash() becomes nm_simple_connection_new_from_hash(), and nm_connection_duplicate() becomes nm_simple_connection_new_clone().
2014-08-13 14:34:29 -04:00
connection = nm_simple_connection_new ();
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
/* Build up the 'connection' setting */
s_con = (NMSettingConnection *) nm_setting_connection_new ();
uuid = nm_utils_uuid_generate ();
if (con_name)
default_name = g_strdup (con_name);
else
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
default_name = nmc_unique_connection_name (connections,
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
get_name_alias (connection_type, NULL, nmc_valid_connection_types));
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
g_object_set (s_con,
NM_SETTING_CONNECTION_ID, default_name,
NM_SETTING_CONNECTION_UUID, uuid,
NM_SETTING_CONNECTION_TYPE, connection_type,
NULL);
g_free (uuid);
g_free (default_name);
nm_connection_add_setting (connection, NM_SETTING (s_con));
/* Initialize the new connection so that it is valid from the start */
editor_init_new_connection (nmc, connection);
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: make Ctrl-C exit nmcli when line is empty *and* we are not in the editor
2014-06-23 13:59:46 +02:00
/* nmcli runs the editor */
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
nmc->nmc_config_mutable.in_editor = TRUE;
cli: make Ctrl-C exit nmcli when line is empty *and* we are not in the editor
2014-06-23 13:59:46 +02:00
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print ("\n");
g_print (_("===| nmcli interactive connection editor |==="));
g_print ("\n\n");
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
if (con)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Editing existing '%s' connection: '%s'"), connection_type, con);
cli: allow editing existing connections in interactive editor So the syntax is now: nmcli con edit [id | uuid | path ] <ID> - for editing an existing connection nmcli con edit [type <new connection type>] [con-name <new connection name>] - for adding a new connection
2013-04-19 21:06:20 +02:00
else
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Adding a new '%s' connection"), connection_type);
g_print ("\n\n");
g_print (_("Type 'help' or '?' for available commands."));
g_print ("\n");
g_print (_("Type 'describe [<setting>.<prop>]' for detailed property description."));
g_print ("\n\n");
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
/* Set global variables for use in TAB completion */
nmc_tab_completion.nmc = nmc;
cli: fix TAB-completion and 'goto' after saving connection On 'save' we replace settings in the local connection, so we also need to update stored setting pointers.
2014-01-20 17:38:11 +01:00
nmc_tab_completion.con_type = g_strdup (connection_type);
cli: TAB-complete devices for 'activate' command in the editor
2013-10-29 12:38:38 +01:00
nmc_tab_completion.connection = connection;
cli: libreadline (libedit) TAB-completion for 'connection edit' TODO: - test with libedit (it seems to be a bit different, buggy?)
2013-04-15 13:35:14 +02:00
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
/* Run menu loop */
editor_menu_main (nmc, connection, connection_type);
if (connection)
g_object_unref (connection);
cli: fix TAB-completion and 'goto' after saving connection On 'save' we replace settings in the local connection, so we also need to update stored setting pointers.
2014-01-20 17:38:11 +01:00
g_free (nmc_tab_completion.con_type);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
return nmc->return_value;
error:
coverity: fix various warnings detected with Coverity These are (most likely) only warnings and not severe bugs. Some of these changes are mostly made to get a clean run of Coverity without any warnings. Error found by running Coverity scan https://bugzilla.redhat.com/show_bug.cgi?id=1025894 Co-Authored-By: Jiří Klimeš <jklimes@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-31 14:13:33 +01:00
g_assert (!connection);
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
g_free (type_ask);
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
return nmc->return_value;
}
cli: 'connection edit' - edit/add connections in an interactive editor nmcli connection edit [type <type>] [con-name <name>] Examples: nmcli c e nmcli c e type ethernet nmcli c e type wifi con-name "My home Wi-Fi" nmcli c e type bond ...
2013-01-18 15:21:00 +01:00
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
static void
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
modify_connection_cb (GObject *connection,
GAsyncResult *result,
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
gpointer user_data)
{
NmCli *nmc = (NmCli *) user_data;
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
GError *error = NULL;
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
if (!nm_remote_connection_commit_changes_finish (NM_REMOTE_CONNECTION (connection),
result, &error)) {
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
g_string_printf (nmc->return_text,
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
_("Error: Failed to modify connection '%s': %s"),
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
nm_connection_get_id (NM_CONNECTION (connection)),
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
error->message);
g_error_free (error);
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
} else {
cli: pass configuration options separately from NmCli structure The NmCli structure is passed around everywhere and contains all the state of the program. It is very hard to follow which parts are used where. Split out more configuration options to a NmcConfig field. This field is mostly immutable, and modified at particular places that now can be easily found.
2017-03-30 12:45:41 +02:00
if (nmc->nmc_config.print_output == NMC_PRINT_PRETTY)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_print (_("Connection '%s' (%s) successfully modified.\n"),
nm_connection_get_id (NM_CONNECTION (connection)),
nm_connection_get_uuid (NM_CONNECTION (connection)));
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
}
quit ();
}
static NMCResultCode
cli: set vs. append property value by 'nmcli con modify' (rh #1044027) Previously 'nmcli con modify' appended values for multi-value properties. This commit makes 'nmcli con modify' overwrite the whole value. You can choose appending values by prefixing the setting.property with '+' sign. For simple (not container) properties the behaviour is the same both with and without the '+', of course. Synopsis: nmcli connection modify [+]<setting>.<property <value> Example: ---> ipv4.dns = 1.2.3.4 $ nmcli connection modify my-em1 ipv4.dns 8.8.8.8 ---> ipv4.dns = 8.8.8.8 $ nmcli connection modify my-em1 +ipv4.dns 8.8.4.4 ---> ipv4.dns = 8.8.8.8 8.8.4.4 https://bugzilla.redhat.com/show_bug.cgi?id=1044027
2014-02-06 11:12:18 +01:00
do_connection_modify (NmCli *nmc,
int argc,
char **argv)
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
{
NMConnection *connection = NULL;
NMRemoteConnection *rc = NULL;
GError *error = NULL;
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
gboolean temporary = FALSE;
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
/* Check --temporary */
if (next_arg (nmc, &argc, &argv, "--temporary", NULL) > 0) {
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
temporary = TRUE;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
connection = get_connection (nmc, &argc, &argv, NULL, &error);
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
if (!connection) {
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_string_printf (nmc->return_text, _("Error: %s."), error->message);
nmc->return_value = error->code;
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
goto finish;
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
rc = nm_client_get_connection_by_uuid (nmc->client,
nm_connection_get_uuid (connection));
cli: allow modifying multiple properties by 'nmcli con modify' This is necessary especially for cases where properties depend on each other. So you need to set them in one command, else the profile won't validate. Examples: nmcli con mod em1-1 ipv4.method manual ipv4.addr "192.168.1.2/24,10.10.1.5/8" nmcli con mod profile ipv4.method link-local ipv4.addr ""
2014-02-12 10:56:10 +01:00
if (!rc) {
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_string_printf (nmc->return_text, _("Error: Unknown connection '%s'."),
nm_connection_get_uuid (connection));
cli: allow modifying multiple properties by 'nmcli con modify' This is necessary especially for cases where properties depend on each other. So you need to set them in one command, else the profile won't validate. Examples: nmcli con mod em1-1 ipv4.method manual ipv4.addr "192.168.1.2/24,10.10.1.5/8" nmcli con mod profile ipv4.method link-local ipv4.addr ""
2014-02-12 10:56:10 +01:00
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
goto finish;
}
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
cli/connection: export read_connection_properties() It makes sense to modify the applied connection from the device object.
2016-06-22 17:35:16 +02:00
if (!nmc_read_connection_properties (nmc, NM_CONNECTION (rc), &argc, &argv, &error)) {
cli: fix error string format clients/cli/.libs/nmcli con mod profile con.zone Error: Error: value for 'con.zone' is missing.. Fixes: 15149d915f9f0d0c7870a01b6f61723296495116
2015-09-25 15:32:31 +02:00
g_string_assign (nmc->return_text, error->message);
cli: refactor: split connection property reader from do_connection_modify() No functional change. We'll need it for specifying the properties on connection addition.
2015-06-07 14:00:57 +02:00
nmc->return_value = error->code;
g_clear_error (&error);
goto finish;
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
}
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
if (nmc->complete)
goto finish;
update_connection (!temporary, rc, modify_connection_cb, nmc);
nmc->should_wait++;
cli: allow modifying multiple properties by 'nmcli con modify' This is necessary especially for cases where properties depend on each other. So you need to set them in one command, else the profile won't validate. Examples: nmcli con mod em1-1 ipv4.method manual ipv4.addr "192.168.1.2/24,10.10.1.5/8" nmcli con mod profile ipv4.method link-local ipv4.addr ""
2014-02-12 10:56:10 +01:00
cli: 'connection modify' - modify a single property nmcli connection modify [id | uuid | path] <ID> <setting>.<property> [<value>] missing value means setting the default value (deleting). Examples: nmcli con mod ethernet-2 connection.autoconnect no nmcli con mod "Home Wi-Fi" wifi.mtu 1350 nmcli con mod "T-Mobile 1" gsm.apn internet nmcli con mod 91782692-512e-4408-9572-667887319ef4 conn.perm user:cimrman TODO: - allow modifying multiple properties (using single command)
2013-04-13 14:33:28 +02:00
finish:
return nmc->return_value;
}
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
typedef struct {
NmCli *nmc;
char *orig_id;
char *orig_uuid;
char *con_id;
} CloneConnectionInfo;
static void
clone_connection_cb (GObject *client,
GAsyncResult *result,
gpointer user_data)
{
CloneConnectionInfo *info = (CloneConnectionInfo *) user_data;
NmCli *nmc = info->nmc;
NMRemoteConnection *connection;
GError *error = NULL;
connection = nm_client_add_connection_finish (NM_CLIENT (client), result, &error);
if (error) {
g_string_printf (nmc->return_text,
_("Error: Failed to add '%s' connection: %s"),
info->con_id, error->message);
g_error_free (error);
nmc->return_value = NMC_RESULT_ERROR_CON_ACTIVATION;
} else {
g_print (_("%s (%s) cloned as %s (%s).\n"),
info->orig_id,
info->orig_uuid,
nm_connection_get_id (NM_CONNECTION (connection)),
nm_connection_get_uuid (NM_CONNECTION (connection)));
g_object_unref (connection);
}
g_free (info->con_id);
g_free (info->orig_id);
g_free (info->orig_uuid);
g_slice_free (CloneConnectionInfo, info);
quit ();
}
static NMCResultCode
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
do_connection_clone (NmCli *nmc, int argc, char **argv)
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
{
NMConnection *connection = NULL;
NMConnection *new_connection = NULL;
NMSettingConnection *s_con;
CloneConnectionInfo *info;
const char *new_name;
char *new_name_ask = NULL;
char *uuid;
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
gboolean temporary = FALSE;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
char **arg_arr = NULL;
int arg_num;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
char ***argv_ptr;
int *argc_ptr;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
GError *error = NULL;
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
if (next_arg (nmc, &argc, &argv, "--temporary", NULL) > 0) {
temporary = TRUE;
next_arg (nmc, &argc, &argv, NULL);
}
argv_ptr = &argv;
argc_ptr = &argc;
nmcli: improve connection autocompletion (2/2) Complete the property as we parse the list of properties. This makes it possible to actually complete an unfinished property. E.g: $ nmcli --complete c modify enp0s25 +ipv6.addr +ipv6.addresses +ipv6.addr-gen-mode
2016-06-15 13:12:39 +02:00
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
if (argc == 0 && nmc->ask) {
cli/connections: add completion to clone subcommand
2016-07-07 14:54:51 +02:00
char *line;
/* nmc_do_cmd() should not call this with argc=0. */
g_assert (!nmc->complete);
line = nmc_readline ("%s: ", PROMPT_CONNECTION);
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
nmc_string_to_arg_array (line, NULL, TRUE, &arg_arr, &arg_num);
g_free (line);
argv_ptr = &arg_arr;
argc_ptr = &arg_num;
cli: fix a crash in 'nmcli -a con clone' when providing empty arguments. Fixes: 0c65b289601d46edb0a950e291ddd376e368ccfd
2015-11-27 20:58:54 +01:00
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
connection = get_connection (nmc, argc_ptr, argv_ptr, NULL, &error);
if (!connection) {
g_string_printf (nmc->return_text, _("Error: %s."), error->message);
nmc->return_value = error->code;
cli: fix a crash in 'nmcli -a con clone' when providing empty arguments. Fixes: 0c65b289601d46edb0a950e291ddd376e368ccfd
2015-11-27 20:58:54 +01:00
goto finish;
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
cli/connections: add completion to clone subcommand
2016-07-07 14:54:51 +02:00
if (nmc->complete)
goto finish;
nmcli: fix connection clone/modify remove extra "next_arg"s (get_connection already shifts argv). Fixes: 717db4fe014ecc80eb8bb86757f724448116a662
2016-08-02 15:11:28 +02:00
if (argv[0])
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
new_name = *argv;
else if (nmc->ask)
new_name = new_name_ask = nmc_readline (_("New connection name: "));
else {
cli: fix a crash in 'nmcli -a con clone' when providing empty arguments. Fixes: 0c65b289601d46edb0a950e291ddd376e368ccfd
2015-11-27 20:58:54 +01:00
g_string_printf (nmc->return_text, _("Error: <new name> argument is missing."));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
}
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
if (next_arg (nmc->ask ? NULL : nmc, argc_ptr, argv_ptr, NULL) == 0) {
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_string_printf (nmc->return_text, _("Error: unknown extra argument: '%s'."), *argv);
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
goto finish;
}
/* Copy the connection */
new_connection = nm_simple_connection_new_clone (connection);
s_con = nm_connection_get_setting_connection (new_connection);
g_assert (s_con);
uuid = nm_utils_uuid_generate ();
g_object_set (s_con,
NM_SETTING_CONNECTION_ID, new_name,
NM_SETTING_CONNECTION_UUID, uuid,
NULL);
g_free (uuid);
/* Merge secrets into the new connection */
update_secrets_in_connection (NM_REMOTE_CONNECTION (connection), new_connection);
info = g_slice_new0 (CloneConnectionInfo);
info->nmc = nmc;
info->orig_id = g_strdup (nm_connection_get_id (connection));
info->orig_uuid = g_strdup (nm_connection_get_uuid (connection));
info->con_id = g_strdup (nm_connection_get_id (new_connection));
/* Add the new cloned connection to NetworkManager */
add_new_connection (!temporary,
nmc->client,
new_connection,
clone_connection_cb,
info);
cli: use should_wait consistently It's a semaphore, not a boolean.
2016-06-22 17:44:39 +02:00
nmc->should_wait++;
cli: add 'nmcli connection clone' for cloning connections (bgo #757627) Synopsis: nmcli connection clone [--temporary] [id|uuid|path] <ID> <new name> It copies the <ID> connection as <new name>. The command is very useful if there is a connection, but another one is needed for a related configuration. One can copy the existing profile and modify it for the new situation. For example: $ nmcli con clone main-eth second-eth $ nmcli con modify second-eth connection.interface-name em4 https://bugzilla.gnome.org/show_bug.cgi?id=757627
2015-11-05 12:03:49 +01:00
finish:
if (new_connection)
g_object_unref (new_connection);
g_free (new_name_ask);
return nmc->return_value;
}
cli: fix 'nmcli con delete' (rh #771258) The command didn't delete the connection, because nmcli exited before NM could check permissions (actually ConsoleKit didn't see the process). Now nmcli waits for "Removed" signal or D-Bus return (error) message on "Delete" method call.
2012-01-10 16:18:27 +01:00
static void
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
delete_cb (GObject *con, GAsyncResult *result, gpointer user_data)
cli: fix 'nmcli con delete' (rh #771258) The command didn't delete the connection, because nmcli exited before NM could check permissions (actually ConsoleKit didn't see the process). Now nmcli waits for "Removed" signal or D-Bus return (error) message on "Delete" method call.
2012-01-10 16:18:27 +01:00
{
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
ConnectionCbInfo *info = (ConnectionCbInfo *) user_data;
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
GError *error = NULL;
cli: fix 'nmcli con delete' (rh #771258) The command didn't delete the connection, because nmcli exited before NM could check permissions (actually ConsoleKit didn't see the process). Now nmcli waits for "Removed" signal or D-Bus return (error) message on "Delete" method call.
2012-01-10 16:18:27 +01:00
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
if (!nm_remote_connection_delete_finish (NM_REMOTE_CONNECTION (con), result, &error)) {
cli: avoid use-after free on connection deletion If the connection spontaneously disappears (perhaps along with the whole daemon on crash) while we're deleting it, then the removal callback would free up the context structure the delete operation is using. Let's cancel the in-flight delete operations so that they won't touch the structure after it's gone.
2017-01-26 15:59:01 +01:00
if (g_error_matches (error, G_IO_ERROR, G_IO_ERROR_CANCELLED))
return;
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
g_string_printf (info->nmc->return_text, _("Error: not all connections deleted."));
g_printerr (_("Error: Connection deletion failed: %s"),
error->message);
libnm: make sync/async APIs more GLib-like Make synchronous APIs take GCancellables, and make asynchronous APIs use GAsyncReadyCallbacks and have names ending in "_async", with "_finish" functions to retrieve the results. Also, make nm_client_activate_connection_finish(), nm_client_add_and_activate_finish(), and nm_remote_settings_add_connection_finish() be (transfer full) rather than (transfer none), because the refcounting semantics become slightly confusing in some edge cases otherwise.
2014-09-11 16:27:13 -04:00
g_error_free (error);
cli: make id|uuid optional for 'nmcli connection delete' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection delete [id|uuid|path] <ID> Example: nmcli connection delete "Wired 1" "Wired 2" "Wired 3" nmcli connection delete id MyMain id "Quick Test 1" "Quick Test 2" path 23 path 47
2012-11-19 13:53:24 +01:00
info->nmc->return_value = NMC_RESULT_ERROR_CON_DEL;
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
connection_cb_info_finish (info, con);
} else {
if (info->nmc->nowait_flag)
connection_cb_info_finish (info, con);
cli: make id|uuid optional for 'nmcli connection delete' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection delete [id|uuid|path] <ID> Example: nmcli connection delete "Wired 1" "Wired 2" "Wired 3" nmcli connection delete id MyMain id "Quick Test 1" "Quick Test 2" path 23 path 47
2012-11-19 13:53:24 +01:00
}
cli: fix 'nmcli con delete' (rh #771258) The command didn't delete the connection, because nmcli exited before NM could check permissions (actually ConsoleKit didn't see the process). Now nmcli waits for "Removed" signal or D-Bus return (error) message on "Delete" method call.
2012-01-10 16:18:27 +01:00
}
cli: add 'nmcli con delete' command for removing configured connections
2011-09-06 15:07:58 +02:00
static NMCResultCode
do_connection_delete (NmCli *nmc, int argc, char **argv)
{
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
NMConnection *connection;
ConnectionCbInfo *info = NULL;
GSList *queue = NULL, *iter;
cli: fix error reporting for connection delete get_connection() shifts to the next argument, so we need to cache the old one. Fixes: 717db4fe014ecc80eb8bb86757f724448116a662
2016-08-02 17:11:50 +02:00
char **arg_arr = NULL, *old_arg;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
char **arg_ptr;
int arg_num;
cli: 'connection delete' - do not stop on invalid connection arguments Rather report it and continue with the next one. This is the behaviour e.g. 'rm' command uses.
2013-04-05 13:01:08 +02:00
GString *invalid_cons = NULL;
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
int pos = 0;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
GError *error = NULL;
cli: add 'nmcli con delete' command for removing configured connections
2011-09-06 15:07:58 +02:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
if (nmc->timeout == -1)
nmc->timeout = 10;
cli: add 'nmcli con delete' command for removing configured connections
2011-09-06 15:07:58 +02:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
arg_ptr = argv;
arg_num = argc;
cli: make id|uuid optional for 'nmcli connection delete' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection delete [id|uuid|path] <ID> Example: nmcli connection delete "Wired 1" "Wired 2" "Wired 3" nmcli connection delete id MyMain id "Quick Test 1" "Quick Test 2" path 23 path 47
2012-11-19 13:53:24 +01:00
if (argc == 0) {
cli: 'connection delete' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:54:41 +01:00
if (nmc->ask) {
cli/connections: add completion to delete subcommand
2016-07-07 14:55:08 +02:00
char *line;
/* nmc_do_cmd() should not call this with argc=0. */
g_assert (!nmc->complete);
line = nmc_readline ("%s: ", PROMPT_CONNECTIONS);
cli: fix TAB completion for multiple connections in 'nmcli con down/delete'
2015-03-12 14:37:47 +01:00
nmc_string_to_arg_array (line, NULL, TRUE, &arg_arr, &arg_num);
cli: fix memory leaks when asking for arguments
2015-03-12 14:42:42 +01:00
g_free (line);
cli: 'connection delete' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:54:41 +01:00
arg_ptr = arg_arr;
}
if (arg_num == 0) {
g_string_printf (nmc->return_text, _("Error: No connection specified."));
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
cli: 'connection delete' - do not stop on invalid connection arguments Rather report it and continue with the next one. This is the behaviour e.g. 'rm' command uses.
2013-04-05 13:01:08 +02:00
goto finish;
cli: 'connection delete' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:54:41 +01:00
}
cli: add 'nmcli con delete' command for removing configured connections
2011-09-06 15:07:58 +02:00
}
cli: 'connection delete' - ask for a connection name if not provided when '--ask' switch is used
2012-11-22 13:54:41 +01:00
while (arg_num > 0) {
cli: fix error reporting for connection delete get_connection() shifts to the next argument, so we need to cache the old one. Fixes: 717db4fe014ecc80eb8bb86757f724448116a662
2016-08-02 17:11:50 +02:00
old_arg = *arg_ptr;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
connection = get_connection (nmc, &arg_num, &arg_ptr, &pos, &error);
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
if (connection) {
/* Check if the connection is unique. */
/* Calling delete for the same connection repeatedly would result in
* NM responding for the last D-Bus call only and we would stall. */
if (!g_slist_find (queue, connection))
queue = g_slist_prepend (queue, g_object_ref (connection));
} else {
cli/connections: add completion to delete subcommand
2016-07-07 14:55:08 +02:00
if (!nmc->complete)
g_printerr (_("Error: %s.\n"), error->message);
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_string_printf (nmc->return_text, _("Error: not all connections found."));
nmc->return_value = error->code;
g_clear_error (&error);
if (nmc->return_value != NMC_RESULT_ERROR_NOT_FOUND)
goto finish;
cli: 'connection delete' - do not stop on invalid connection arguments Rather report it and continue with the next one. This is the behaviour e.g. 'rm' command uses.
2013-04-05 13:01:08 +02:00
if (!invalid_cons)
invalid_cons = g_string_new (NULL);
cli: fix error reporting for connection delete get_connection() shifts to the next argument, so we need to cache the old one. Fixes: 717db4fe014ecc80eb8bb86757f724448116a662
2016-08-02 17:11:50 +02:00
g_string_append_printf (invalid_cons, "'%s', ", old_arg);
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
}
}
cli: make id|uuid optional for 'nmcli connection delete' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection delete [id|uuid|path] <ID> Example: nmcli connection delete "Wired 1" "Wired 2" "Wired 3" nmcli connection delete id MyMain id "Quick Test 1" "Quick Test 2" path 23 path 47
2012-11-19 13:53:24 +01:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
if (!queue) {
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_string_printf (nmc->return_text, _("Error: No connection specified."));
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
goto finish;
cli/connections: add completion to delete subcommand
2016-07-07 14:55:08 +02:00
} else if (nmc->complete) {
g_slist_free (queue);
goto finish;
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
}
queue = g_slist_reverse (queue);
cli: make id|uuid optional for 'nmcli connection delete' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection delete [id|uuid|path] <ID> Example: nmcli connection delete "Wired 1" "Wired 2" "Wired 3" nmcli connection delete id MyMain id "Quick Test 1" "Quick Test 2" path 23 path 47
2012-11-19 13:53:24 +01:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
info = g_slice_new0 (ConnectionCbInfo);
info->nmc = nmc;
info->queue = queue;
info->timeout_id = g_timeout_add_seconds (nmc->timeout, connection_op_timeout_cb, info);
cli: avoid use-after free on connection deletion If the connection spontaneously disappears (perhaps along with the whole daemon on crash) while we're deleting it, then the removal callback would free up the context structure the delete operation is using. Let's cancel the in-flight delete operations so that they won't touch the structure after it's gone.
2017-01-26 15:59:01 +01:00
info->cancellable = g_cancellable_new ();
cli: 'connection delete' - do not stop on invalid connection arguments Rather report it and continue with the next one. This is the behaviour e.g. 'rm' command uses.
2013-04-05 13:01:08 +02:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
nmc->nowait_flag = (nmc->timeout == 0);
cli: turn should_wait into a semaphore Count the exit blockers. We'll want to terminate the device monitor in case all monitored devices vanish.
2015-03-27 12:39:17 +01:00
nmc->should_wait++;
cli: make id|uuid optional for 'nmcli connection delete' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection delete [id|uuid|path] <ID> Example: nmcli connection delete "Wired 1" "Wired 2" "Wired 3" nmcli connection delete id MyMain id "Quick Test 1" "Quick Test 2" path 23 path 47
2012-11-19 13:53:24 +01:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
g_signal_connect (nmc->client, NM_CLIENT_CONNECTION_REMOVED,
G_CALLBACK (connection_removed_cb), info);
cli: make id|uuid optional for 'nmcli connection delete' and also allow identifying connetions with 'path' that accepts either the whole D-Dus path or just an index. nmcli connection delete [id|uuid|path] <ID> Example: nmcli connection delete "Wired 1" "Wired 2" "Wired 3" nmcli connection delete id MyMain id "Quick Test 1" "Quick Test 2" path 23 path 47
2012-11-19 13:53:24 +01:00
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
/* Now delete the connections */
for (iter = queue; iter; iter = g_slist_next (iter))
nm_remote_connection_delete_async (NM_REMOTE_CONNECTION (iter->data),
cli: avoid use-after free on connection deletion If the connection spontaneously disappears (perhaps along with the whole daemon on crash) while we're deleting it, then the removal callback would free up the context structure the delete operation is using. Let's cancel the in-flight delete operations so that they won't touch the structure after it's gone.
2017-01-26 15:59:01 +01:00
info->cancellable, delete_cb, info);
cli: fix 'nmcli con delete' (rh #771258) The command didn't delete the connection, because nmcli exited before NM could check permissions (actually ConsoleKit didn't see the process). Now nmcli waits for "Removed" signal or D-Bus return (error) message on "Delete" method call.
2012-01-10 16:18:27 +01:00
cli: 'connection delete' - do not stop on invalid connection arguments Rather report it and continue with the next one. This is the behaviour e.g. 'rm' command uses.
2013-04-05 13:01:08 +02:00
finish:
if (invalid_cons) {
g_string_truncate (invalid_cons, invalid_cons->len-2); /* truncate trailing ", " */
g_string_printf (nmc->return_text, _("Error: cannot delete unknown connection(s): %s."),
invalid_cons->str);
cli: add NMC_RESULT_ERROR_NOT_FOUND and use as appropriate This will allow callers to distinguish "no such connection" from "connection failed to activate", etc. https://bugzilla.redhat.com/show_bug.cgi?id=1022256
2013-11-21 09:19:18 -05:00
nmc->return_value = NMC_RESULT_ERROR_NOT_FOUND;
cli: 'connection delete' - do not stop on invalid connection arguments Rather report it and continue with the next one. This is the behaviour e.g. 'rm' command uses.
2013-04-05 13:01:08 +02:00
g_string_free (invalid_cons, TRUE);
}
cli: do not stall in 'nmcli connection delete/down' (rh #1168657) NetworkManager only responds to the last D-Bus call when called delete/down for the same connection in quick succession. (It should be fixed later). So do not issue the call multiple times to prevent that. Otherwise nmcli would stall waiting for the response. https://bugzilla.redhat.com/show_bug.cgi?id=1168657
2015-03-12 17:21:59 +01:00
g_strfreev (arg_arr);
cli: add 'nmcli con delete' command for removing configured connections
2011-09-06 15:07:58 +02:00
return nmc->return_value;
}
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
static void
connection_changed (NMConnection *connection, NmCli *nmc)
{
g_print (_("%s: connection profile changed\n"), nm_connection_get_id (connection));
}
static void
connection_watch (NmCli *nmc, NMConnection *connection)
{
nmc->should_wait++;
g_signal_connect (connection, NM_CONNECTION_CHANGED, G_CALLBACK (connection_changed), nmc);
}
static void
connection_unwatch (NmCli *nmc, NMConnection *connection)
{
if (g_signal_handlers_disconnect_by_func (connection, G_CALLBACK (connection_changed), nmc))
nmc->should_wait--;
/* Terminate if all the watched connections disappeared. */
if (!nmc->should_wait)
quit ();
}
static void
connection_added (NMClient *client, NMRemoteConnection *con, NmCli *nmc)
{
NMConnection *connection = NM_CONNECTION (con);
g_print (_("%s: connection profile created\n"), nm_connection_get_id (connection));
connection_watch (nmc, connection);
}
static void
connection_removed (NMClient *client, NMRemoteConnection *con, NmCli *nmc)
{
NMConnection *connection = NM_CONNECTION (con);
g_print (_("%s: connection profile removed\n"), nm_connection_get_id (connection));
connection_unwatch (nmc, connection);
}
static NMCResultCode
do_connection_monitor (NmCli *nmc, int argc, char **argv)
{
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
GError *error = NULL;
nmcli: improve connection autocompletion (2/2) Complete the property as we parse the list of properties. This makes it possible to actually complete an unfinished property. E.g: $ nmcli --complete c modify enp0s25 +ipv6.addr +ipv6.addresses +ipv6.addr-gen-mode
2016-06-15 13:12:39 +02:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
if (argc == 0) {
/* No connections specified. Monitor all. */
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
int i;
cli/connections: add completion to monitor subcommand
2016-07-07 15:08:00 +02:00
/* nmc_do_cmd() should not call this with argc=0. */
g_assert (!nmc->complete);
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nmc->client);
for (i = 0; i < connections->len; i++)
connection_watch (nmc, g_ptr_array_index (connections, i));
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
/* We'll watch the connection additions too, never exit. */
nmc->should_wait++;
g_signal_connect (nmc->client, NM_CLIENT_CONNECTION_ADDED, G_CALLBACK (connection_added), nmc);
} else {
/* Look up the specified connections and watch them. */
NMConnection *connection;
int pos = 0;
do {
cli/connections: add completion to monitor subcommand
2016-07-07 15:08:00 +02:00
connection = get_connection (nmc, &argc, &argv, &pos, &error);
if (!connection) {
if (!nmc->complete)
g_printerr (_("Error: %s.\n"), error->message);
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
g_string_printf (nmc->return_text, _("Error: not all connections found."));
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
return error->code;
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
}
cli/connections: add completion to monitor subcommand
2016-07-07 15:08:00 +02:00
if (nmc->complete)
continue;
connection_watch (nmc, connection);
} while (argc > 0);
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
}
cli/connections: add completion to monitor subcommand
2016-07-07 15:08:00 +02:00
if (nmc->complete)
return nmc->return_value;
cli,connections: add connection monitor
2015-03-27 14:59:38 +01:00
g_signal_connect (nmc->client, NM_CLIENT_CONNECTION_REMOVED, G_CALLBACK (connection_removed), nmc);
return NMC_RESULT_SUCCESS;
}
core: add monitor-connection-files=false and ReloadConnections Add a "monitor-connection-files" config option, which can be set to "false" to disable automatic reloading of connections on file change. To go with this, add a new ReloadConnections method on o.fd.NM.Settings that can be used to manually reload connections, and add an nm-cli command to call it.
2013-05-23 19:05:40 -03:00
static NMCResultCode
do_connection_reload (NmCli *nmc, int argc, char **argv)
{
GError *error = NULL;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
nmcli: improve connection autocompletion (2/2) Complete the property as we parse the list of properties. This makes it possible to actually complete an unfinished property. E.g: $ nmcli --complete c modify enp0s25 +ipv6.addr +ipv6.addresses +ipv6.addr-gen-mode
2016-06-15 13:12:39 +02:00
if (nmc->complete)
return nmc->return_value;
core: add monitor-connection-files=false and ReloadConnections Add a "monitor-connection-files" config option, which can be set to "false" to disable automatic reloading of connections on file change. To go with this, add a new ReloadConnections method on o.fd.NM.Settings that can be used to manually reload connections, and add an nm-cli command to call it.
2013-05-23 19:05:40 -03:00
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
if (!nm_client_reload_connections (nmc->client, NULL, &error)) {
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
g_string_printf (nmc->return_text, _("Error: failed to reload connections: %s."),
cli: return sane error message for D-Bus policy permission errors The error returned to users when a load_connection(s)/set_logging call fails due to D-Bus policy denial is a bit obscure: $ nmcli general logging level debug Error: failed to set logging: Rejected send message, 4 matched rules; type="method_call", sender=":1.233" (uid=1001 pid=27225 comm="nmcli general logging level debug ") interface="org.freedesktop.NetworkManager" member="SetLogging" error name="(unset)" requested_reply="0" destination=":1.207" (uid=0 pid=25793 comm="/usr/sbin/NetworkManager --no-daemon ") Convert it to a more comprehensible: $ nmcli general logging level debug Error: failed to set logging: access denied https://bugzilla.redhat.com/show_bug.cgi?id=1362542 (cherry picked from commit 805925f9efbc910975104bebe1050f0c663b61f7)
2016-08-19 11:56:58 +02:00
nmc_error_get_simple_message (error));
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
core: add monitor-connection-files=false and ReloadConnections Add a "monitor-connection-files" config option, which can be set to "false" to disable automatic reloading of connections on file change. To go with this, add a new ReloadConnections method on o.fd.NM.Settings that can be used to manually reload connections, and add an nm-cli command to call it.
2013-05-23 19:05:40 -03:00
g_clear_error (&error);
}
return nmc->return_value;
}
cli: add "nmcli con load file..." https://bugzilla.gnome.org/show_bug.cgi?id=709830
2013-10-31 15:18:47 -04:00
static NMCResultCode
do_connection_load (NmCli *nmc, int argc, char **argv)
{
GError *error = NULL;
char **filenames, **failures = NULL;
int i;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli: add "nmcli con load file..." https://bugzilla.gnome.org/show_bug.cgi?id=709830
2013-10-31 15:18:47 -04:00
if (argc == 0) {
g_string_printf (nmc->return_text, _("Error: No connection specified."));
cli/connections: add completion to load subcommand
2016-07-12 00:21:55 +02:00
return NMC_RESULT_ERROR_USER_INPUT;
cli: add "nmcli con load file..." https://bugzilla.gnome.org/show_bug.cgi?id=709830
2013-10-31 15:18:47 -04:00
}
cli/connections: add completion to load subcommand
2016-07-12 00:21:55 +02:00
if (nmc->complete)
return NMC_RESULT_COMPLETE_FILE;
cli: add "nmcli con load file..." https://bugzilla.gnome.org/show_bug.cgi?id=709830
2013-10-31 15:18:47 -04:00
filenames = g_new (char *, argc + 1);
for (i = 0; i < argc; i++)
filenames[i] = argv[i];
filenames[i] = NULL;
libnm: merge NMRemoteSettings into NMClient Make NMRemoteSettings internal and have NMClient wrap all of its APIs, just like it does with NMManager.
2014-09-29 10:58:16 -04:00
nm_client_load_connections (nmc->client, filenames, &failures, NULL, &error);
cli: add "nmcli con load file..." https://bugzilla.gnome.org/show_bug.cgi?id=709830
2013-10-31 15:18:47 -04:00
g_free (filenames);
if (error) {
cli: strip errors from D-Bus before presenting to a user
2014-10-01 10:31:03 +02:00
g_string_printf (nmc->return_text, _("Error: failed to load connection: %s."),
cli: return sane error message for D-Bus policy permission errors The error returned to users when a load_connection(s)/set_logging call fails due to D-Bus policy denial is a bit obscure: $ nmcli general logging level debug Error: failed to set logging: Rejected send message, 4 matched rules; type="method_call", sender=":1.233" (uid=1001 pid=27225 comm="nmcli general logging level debug ") interface="org.freedesktop.NetworkManager" member="SetLogging" error name="(unset)" requested_reply="0" destination=":1.207" (uid=0 pid=25793 comm="/usr/sbin/NetworkManager --no-daemon ") Convert it to a more comprehensible: $ nmcli general logging level debug Error: failed to set logging: access denied https://bugzilla.redhat.com/show_bug.cgi?id=1362542 (cherry picked from commit 805925f9efbc910975104bebe1050f0c663b61f7)
2016-08-19 11:56:58 +02:00
nmc_error_get_simple_message (error));
cli: add "nmcli con load file..." https://bugzilla.gnome.org/show_bug.cgi?id=709830
2013-10-31 15:18:47 -04:00
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
g_error_free (error);
}
if (failures) {
for (i = 0; failures[i]; i++)
cli: use g_print/g_printerr rather than printf/fprintf Use g_print() and g_printerr(), which wrap printf() and fprintf(stderr,...), but handle conversion from UTF-8 if the locale is using a different character set.
2014-09-19 16:04:40 -04:00
g_printerr (_("Could not load file '%s'\n"), failures[i]);
cli: add "nmcli con load file..." https://bugzilla.gnome.org/show_bug.cgi?id=709830
2013-10-31 15:18:47 -04:00
g_strfreev (failures);
}
return nmc->return_value;
}
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
// FIXME: change the text when non-VPN connection types are supported
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
#define PROMPT_IMPORT_TYPE PROMPT_VPN_TYPE
#define PROMPT_IMPORT_FILE N_("File to import: ")
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
cli: complete VPN types
2017-02-14 15:01:48 +01:00
static void
nmc_complete_vpn_service (const char *prefix)
{
char **services;
char **candidate;
services = nm_vpn_plugin_info_list_get_service_types (NULL, FALSE, TRUE);
for (candidate = services; *candidate; candidate++) {
if (!*prefix && g_str_has_prefix (*candidate, NM_DBUS_INTERFACE))
continue;
cli: make match() return boolean Coccinelle semantic patch: @@ @@ -int +gboolean matches (...); @@ expression pattern, cmd, len; @@ -int +gboolean matches (...) { ... - return memcmp (pattern, cmd, len); + return memcmp (pattern, cmd, len) == 0; } @@ expression prefix, str; @@ ( -matches (prefix, str) != 0 +!matches (prefix, str) | -matches (prefix, str) == 0 +matches (prefix, str) ) @@ expression prefix, str; @@ -(matches (prefix, str)) +matches (prefix, str) @@ expression prefix, str; @@ -(!matches (prefix, str)) +!matches (prefix, str) spatch --smpl-spacing --sp-file match.cocci --dir clients/cli/ \ --include-headers --macro-file shared/nm-utils/gsystem-local-alloc.h
2017-02-15 12:20:55 +01:00
if (!*prefix || matches (prefix, *candidate))
cli: complete VPN types
2017-02-14 15:01:48 +01:00
g_print ("%s\n", *candidate);
}
g_strfreev (services);
}
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
static NMCResultCode
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
do_connection_import (NmCli *nmc, int argc, char **argv)
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
{
GError *error = NULL;
const char *type = NULL, *filename = NULL;
char *type_ask = NULL, *filename_ask = NULL;
AddConnectionInfo *info;
NMConnection *connection = NULL;
NMVpnEditorPlugin *plugin;
cli: fix handling of VPN names in nmcli by using libnm functions At various places, nmcli requires to specify a VPN type by name, for example $ nmcli connection add type vpn ifname '*' vpn-type $VPN_TYPE This $VPN_TYPE used to be a hard-coded list of known VPN plugin names. But actually, it should be a VPN service-type. A service-type used to be the D-Bus name of the VPN plugin. Now, with multiple VPN support that is no longer the case, but it still has the form of a D-Bus bus name. Alternativley, it could be an alias, which is just a way for plugins to support multiple service-types. Fix that, to support fully qualified service-types in the form of D-Bus bus names. Also, support lookup by name, in which case the present plugin-info instances are searched. Finally, support a list of hard-code short-names. All the logic how to translate a short-name to a fully qualified service-type is now inside libnm, so that various user agree on those names and don't have to hard-code them each.
2016-06-07 11:06:14 +02:00
gs_free char *service_type = NULL;
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
gboolean temporary = FALSE;
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
/* Check --temporary */
if (next_arg (nmc, &argc, &argv, "--temporary", NULL) > 0) {
temporary = TRUE;
next_arg (nmc, &argc, &argv, NULL);
}
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
if (argc == 0) {
cli/connections: add completion to import subcommand
2016-07-12 00:22:04 +02:00
/* nmc_do_cmd() should not call this with argc=0. */
g_assert (!nmc->complete);
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
if (nmc->ask) {
nmcli: streamline connection addition This is a huge refactoring in attempt to 1.) reduce the horrible redundancy in the connection addition path and 2.) reduce confusion between various sources of property value (command line, properties, interactive mode). * The conversions from the strings was done all over the place: settings.c already does for all sensible properties. The rest is removed. * The validations are done randomly and redundantly: server does some validation, and per-property client validations useful for interactive mode are done in settings.c The rest is removed. * The information about defaults and required options was redundantly scattered in per-type completion functions and interactive mode questionnaries. This is now driven by the option_info[] table. In general, we do our best to just map the command line options to properties and allow mixing them. For the rest there's the check_and_set() callbacks (basically to keep compatibility with previous nmcli versions). This this is now all possible: $ nmcli c add type ethernet ifname '*' This always worked $ nmcli c add type bond-slave save no -- connection.autoconnect no The "save" and "--" still work $ nmcli c add connection.type ethernet ifname eth0 Properties can now be used $ nmcli c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever There's no implementation mandated order of the properties (the type still must be known to determine which properties make sense) $ nmcli --ask c add type ethernet ip4 1.2.3.4 mac 80:86:66:77:88:99 con-name whatever The interactive mode asks only for properties that weren't specified on command line
2016-05-20 20:43:58 +02:00
type_ask = nmc_readline (gettext (PROMPT_IMPORT_TYPE));
filename_ask = nmc_readline (gettext (PROMPT_IMPORT_FILE));
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
type = type_ask = type_ask ? g_strstrip (type_ask) : NULL;
filename = filename_ask = filename_ask ? g_strstrip (filename_ask) : NULL;
} else {
g_string_printf (nmc->return_text, _("Error: No arguments provided."));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
}
while (argc > 0) {
cli/connections: add completion to import subcommand
2016-07-12 00:22:04 +02:00
if (argc == 1 && nmc->complete)
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
nmc_complete_strings (*argv, "type", "file", NULL);
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
if (strcmp (*argv, "type") == 0) {
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
g_string_printf (nmc->return_text, _("Error: %s argument is missing."), *(argv-1));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
cli: complete VPN types
2017-02-14 15:01:48 +01:00
if (argc == 1 && nmc->complete)
nmc_complete_vpn_service (*argv);
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
if (!type)
type = *argv;
else
g_printerr (_("Warning: 'type' already specified, ignoring extra one.\n"));
} else if (strcmp (*argv, "file") == 0) {
nmcli: fix nmcli parsing Fixes the parsing for correct and incorrect commands. e.g.: nmcli connection down id <conn> nmcli -t ... Fixes: 16902a2be66866aa541bb56ccb98c765ac21d1a2
2017-03-29 12:02:14 +02:00
argc--;
argv++;
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (!argc) {
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
g_string_printf (nmc->return_text, _("Error: %s argument is missing."), *(argv-1));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
cli/connections: add completion to import subcommand
2016-07-12 00:22:04 +02:00
if (argc == 1 && nmc->complete)
nmc->return_value = NMC_RESULT_COMPLETE_FILE;
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
if (!filename)
filename = *argv;
else
g_printerr (_("Warning: 'file' already specified, ignoring extra one.\n"));
} else {
cli: drop an extra newline
2015-12-07 11:23:39 +01:00
g_string_printf (nmc->return_text, _("Unknown parameter: %s"), *argv);
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
}
cli/connections: add completion to import subcommand
2016-07-12 00:22:04 +02:00
if (nmc->complete)
goto finish;
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
if (!type) {
g_string_printf (nmc->return_text, _("Error: 'type' argument is required."));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
if (!filename) {
g_string_printf (nmc->return_text, _("Error: 'file' argument is required."));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
cli: fix handling of VPN names in nmcli by using libnm functions At various places, nmcli requires to specify a VPN type by name, for example $ nmcli connection add type vpn ifname '*' vpn-type $VPN_TYPE This $VPN_TYPE used to be a hard-coded list of known VPN plugin names. But actually, it should be a VPN service-type. A service-type used to be the D-Bus name of the VPN plugin. Now, with multiple VPN support that is no longer the case, but it still has the form of a D-Bus bus name. Alternativley, it could be an alias, which is just a way for plugins to support multiple service-types. Fix that, to support fully qualified service-types in the form of D-Bus bus names. Also, support lookup by name, in which case the present plugin-info instances are searched. Finally, support a list of hard-code short-names. All the logic how to translate a short-name to a fully qualified service-type is now inside libnm, so that various user agree on those names and don't have to hard-code them each.
2016-06-07 11:06:14 +02:00
service_type = nm_vpn_plugin_info_list_find_service_type (nm_vpn_get_plugin_infos (), type);
if (!service_type) {
g_string_printf (nmc->return_text, _("Error: failed to find VPN plugin for %s."), type);
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
goto finish;
}
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
/* Import VPN configuration */
cli: fix handling of VPN names in nmcli by using libnm functions At various places, nmcli requires to specify a VPN type by name, for example $ nmcli connection add type vpn ifname '*' vpn-type $VPN_TYPE This $VPN_TYPE used to be a hard-coded list of known VPN plugin names. But actually, it should be a VPN service-type. A service-type used to be the D-Bus name of the VPN plugin. Now, with multiple VPN support that is no longer the case, but it still has the form of a D-Bus bus name. Alternativley, it could be an alias, which is just a way for plugins to support multiple service-types. Fix that, to support fully qualified service-types in the form of D-Bus bus names. Also, support lookup by name, in which case the present plugin-info instances are searched. Finally, support a list of hard-code short-names. All the logic how to translate a short-name to a fully qualified service-type is now inside libnm, so that various user agree on those names and don't have to hard-code them each.
2016-06-07 11:06:14 +02:00
plugin = nm_vpn_get_editor_plugin (service_type, &error);
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
if (!plugin) {
g_string_printf (nmc->return_text, _("Error: failed to load VPN plugin: %s."),
error->message);
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
goto finish;
}
connection = nm_vpn_editor_plugin_import (plugin, filename, &error);
if (!connection) {
g_string_printf (nmc->return_text, _("Error: failed to import '%s': %s."),
filename, error->message);
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
goto finish;
}
info = g_malloc0 (sizeof (AddConnectionInfo));
info->nmc = nmc;
info->con_name = g_strdup (nm_connection_get_id (connection));
/* Add the new imported connection to NetworkManager */
add_new_connection (!temporary,
nmc->client,
connection,
add_connection_cb,
info);
cli: use should_wait consistently It's a semaphore, not a boolean.
2016-06-22 17:44:39 +02:00
nmc->should_wait++;
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
finish:
if (connection)
g_object_unref (connection);
g_clear_error (&error);
g_free (type_ask);
g_free (filename_ask);
return nmc->return_value;
}
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
static NMCResultCode
do_connection_export (NmCli *nmc, int argc, char **argv)
{
NMConnection *connection = NULL;
const char *out_name = NULL;
char *name_ask = NULL;
char *out_name_ask = NULL;
const char *path = NULL;
const char *type = NULL;
NMVpnEditorPlugin *plugin;
GError *error = NULL;
cli: avoid using tmp template after it goes out of scope Fixes: 3892b839af4597c29ebff6d77666566dc4f94fb1
2015-12-07 11:05:50 +01:00
char tmpfile[] = "/tmp/nmcli-export-temp-XXXXXX";
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
char **arg_arr = NULL;
int arg_num;
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
char ***argv_ptr;
int *argc_ptr;
next_arg (nmc, &argc, &argv, NULL);
argv_ptr = &argv;
argc_ptr = &argc;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
if (argc == 0 && nmc->ask) {
cli/connections: add completion to export subcommand
2016-07-07 14:56:22 +02:00
char *line;
/* nmc_do_cmd() should not call this with argc=0. */
g_assert (!nmc->complete);
line = nmc_readline ("%s: ", PROMPT_VPN_CONNECTION);
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
nmc_string_to_arg_array (line, NULL, TRUE, &arg_arr, &arg_num);
g_free (line);
argv_ptr = &arg_arr;
argc_ptr = &arg_num;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
connection = get_connection (nmc, argc_ptr, argv_ptr, NULL, &error);
if (!connection) {
g_string_printf (nmc->return_text, _("Error: %s."), error->message);
nmc->return_value = error->code;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
goto finish;
}
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
cli/connections: add completion to export subcommand
2016-07-07 14:56:22 +02:00
if (nmc->complete)
return nmc->return_value;
nmcli: fix 'nmcli con export' parameters check get_connection() will already move forward arguments (argc/argv): remove extra argv++/argc-- Example: "nmcli con export <vpn_con> <output_file> <extra_arg>" now, extra_arg is detected, printing error: "Error: unknown extra argument: 'extra_arg'."
2017-04-03 12:41:55 +02:00
out_name = *argv;
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
if (next_arg (nmc->ask ? NULL : nmc, argc_ptr, argv_ptr, NULL) == 0) {
cli/connections: split connection name parsing Reduces redundant code paths; will make it easier to implement completion in a common one.
2016-07-06 20:49:40 +02:00
g_string_printf (nmc->return_text, _("Error: unknown extra argument: '%s'."), *argv);
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
goto finish;
}
cli: use next_arg() on argument boundary Shift argc and argc manually between argument ant its value and use next_arg() between arguments everywhere. Whill be useful to parse global arguments.
2017-03-24 14:00:25 +01:00
if (out_name == NULL && nmc->ask)
out_name = out_name_ask = nmc_readline (_("Output file name: "));
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
type = nm_connection_get_connection_type (connection);
if (g_strcmp0 (type, NM_SETTING_VPN_SETTING_NAME) != 0) {
g_string_printf (nmc->return_text, _("Error: the connection is not VPN."));
nmc->return_value = NMC_RESULT_ERROR_USER_INPUT;
goto finish;
}
type = nm_setting_vpn_get_service_type (nm_connection_get_setting_vpn (connection));
/* Export VPN configuration */
cli: fix handling of VPN names in nmcli by using libnm functions At various places, nmcli requires to specify a VPN type by name, for example $ nmcli connection add type vpn ifname '*' vpn-type $VPN_TYPE This $VPN_TYPE used to be a hard-coded list of known VPN plugin names. But actually, it should be a VPN service-type. A service-type used to be the D-Bus name of the VPN plugin. Now, with multiple VPN support that is no longer the case, but it still has the form of a D-Bus bus name. Alternativley, it could be an alias, which is just a way for plugins to support multiple service-types. Fix that, to support fully qualified service-types in the form of D-Bus bus names. Also, support lookup by name, in which case the present plugin-info instances are searched. Finally, support a list of hard-code short-names. All the logic how to translate a short-name to a fully qualified service-type is now inside libnm, so that various user agree on those names and don't have to hard-code them each.
2016-06-07 11:06:14 +02:00
plugin = nm_vpn_get_editor_plugin (type, &error);
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
if (!plugin) {
cli: improve the error reporting a bit
2015-12-07 10:51:49 +01:00
g_string_printf (nmc->return_text, _("Error: failed to load VPN plugin: %s."),
error->message);
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
goto finish;
}
if (out_name)
path = out_name;
else {
int fd;
fd = g_mkstemp (tmpfile);
if (fd == -1) {
g_string_printf (nmc->return_text, _("Error: failed to create temporary file %s."), tmpfile);
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
goto finish;
}
close (fd);
path = tmpfile;
}
if (!nm_vpn_editor_plugin_export (plugin, path, connection, &error)) {
g_string_printf (nmc->return_text, _("Error: failed to export '%s': %s."),
all: clean-up usage of GError Functions that take a GError** MUST fill it in on error. There is no need to check whether error is NULL if the function it was passed to had a failing return value. Likewise, a proper GError must have a non-NULL message, so there's no need to double-check that either. Based-on-patch-by: Dan Winship <danw@gnome.org>
2016-02-28 16:25:36 +01:00
nm_connection_get_id (connection), error->message);
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
goto finish;
}
/* No output file -> copy data to stdout */
if (!out_name) {
char *contents = NULL;
gsize len = 0;
if (!g_file_get_contents (path, &contents, &len, &error)) {
g_string_printf (nmc->return_text, _("Error: failed to read temporary file '%s': %s."),
path, error->message);
nmc->return_value = NMC_RESULT_ERROR_UNKNOWN;
goto finish;
}
g_print ("%s", contents);
g_free (contents);
}
finish:
if (!out_name && path)
unlink (path);
g_clear_error (&error);
g_free (name_ask);
g_free (out_name_ask);
return nmc->return_value;
}
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
static char *
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
gen_func_connection_names (const char *text, int state)
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
{
libnm: change GSList to GPtrArray in libnm methods libnm mostly used GPtrArrays in its APIs, except that arrays of connections were usually GSLists. Fix this and make them GPtrArrays too (and rename nm_client_list_connections() to nm_client_get_connections() to match everything else).
2014-10-22 12:32:46 -04:00
int i;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
const GPtrArray *connections;
const char **connection_names;
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
char *ret;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connections = nm_client_get_connections (nm_cli.client);
if (connections->len == 0)
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
return NULL;
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
connection_names = g_new (const char *, connections->len + 1);
for (i = 0; i < connections->len; i++)
connection_names[i] = nm_connection_get_id (NM_CONNECTION (connections->pdata[i]));
connection_names[i] = NULL;
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
ret = nmc_rl_gen_func_basic (text, state, connection_names);
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
cli: get rid of client-global connections list Caching it in the NmCli object is unnecessary, ugly and would be cumbersome in future when we'll be creating the client object only when needed.
2016-08-31 22:50:02 +02:00
g_free (connection_names);
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
return ret;
}
cli: only offer active connections in TAB completion for 'nmcli con down'
2015-03-13 10:02:37 +01:00
static char *
gen_func_active_connection_names (const char *text, int state)
{
int i;
const GPtrArray *acs;
const char **connections;
char *ret;
if (!nm_cli.client)
return NULL;
acs = nm_client_get_active_connections (nm_cli.client);
if (!acs || acs->len == 0)
return NULL;
connections = g_new (const char *, acs->len + 1);
for (i = 0; i < acs->len; i++)
connections[i] = nm_active_connection_get_id (acs->pdata[i]);
connections[i] = NULL;
ret = nmc_rl_gen_func_basic (text, state, connections);
g_free (connections);
return ret;
}
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
static char **
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
nmcli_con_tab_completion (const char *text, int start, int end)
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
{
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
char **match_array = NULL;
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
rl_compentry_func_t *generator_func = NULL;
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
/* Disable readline's default filename completion */
rl_attempted_completion_over = 1;
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
cli: fix TAB completion for multiple connections in 'nmcli con down/delete'
2015-03-12 14:37:47 +01:00
if (g_strcmp0 (rl_prompt, PROMPT_CONNECTION) == 0) {
/* Disable appending space after completion */
rl_completion_append_character = '\0';
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
cli: fix TAB completion for multiple connections in 'nmcli con down/delete'
2015-03-12 14:37:47 +01:00
if (!is_single_word (rl_line_buffer))
return NULL;
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
generator_func = gen_func_connection_names;
cli: fix TAB completion for multiple connections in 'nmcli con down/delete'
2015-03-12 14:37:47 +01:00
} else if (g_strcmp0 (rl_prompt, PROMPT_CONNECTIONS) == 0) {
generator_func = gen_func_connection_names;
cli: only offer active connections in TAB completion for 'nmcli con down'
2015-03-13 10:02:37 +01:00
} else if (g_strcmp0 (rl_prompt, PROMPT_ACTIVE_CONNECTIONS) == 0) {
generator_func = gen_func_active_connection_names;
cli: add 'nmcli connection import' (rh #1034105) Synopsis: nmcli connection import [--temporary] type <type> file <file to import> Only VPN configurations can be imported at the moment. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-23 17:21:02 +01:00
} else if (g_strcmp0 (rl_prompt, PROMPT_IMPORT_TYPE) == 0) {
generator_func = gen_func_vpn_types;
} else if (g_strcmp0 (rl_prompt, PROMPT_IMPORT_FILE) == 0) {
rl_attempted_completion_over = 0;
rl_complete_with_tilde_expansion = 1;
cli: add 'nmcli connection export' (rh #1034105) Synopsis: nmcli connection export [ id | uuid | path] <ID> [<output file>] for exporting VPN connections. https://bugzilla.redhat.com/show_bug.cgi?id=1034105
2015-11-24 14:14:53 +01:00
} else if (g_strcmp0 (rl_prompt, PROMPT_VPN_CONNECTION) == 0) {
generator_func = gen_vpn_ids;
cli: fix TAB completion for multiple connections in 'nmcli con down/delete'
2015-03-12 14:37:47 +01:00
}
cli: implement TAB completion for connection and interface names For commands like nmcli -a dev disconnect nmcli -a con up ...
2014-05-15 12:52:26 +02:00
if (generator_func)
match_array = rl_completion_matches (text, generator_func);
return match_array;
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
}
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
static const NMCCommand connection_cmds[] = {
cli: use nmc_do_cmd to get the client and check if the daemon is running The makes use of asynchronous client initialization, making things a bit faster and reduces code duplication too.
2016-08-31 20:52:48 +02:00
{ "show", do_connections_show, usage_connection_show, TRUE, TRUE },
{ "up", do_connection_up, usage_connection_up, TRUE, TRUE },
{ "down", do_connection_down, usage_connection_down, TRUE, TRUE },
{ "add", do_connection_add, usage_connection_add, TRUE, TRUE },
{ "edit", do_connection_edit, usage_connection_edit, TRUE, TRUE },
{ "delete", do_connection_delete, usage_connection_delete, TRUE, TRUE },
{ "reload", do_connection_reload, usage_connection_reload, TRUE, TRUE },
{ "load", do_connection_load, usage_connection_load, TRUE, TRUE },
{ "modify", do_connection_modify, usage_connection_modify, TRUE, TRUE },
{ "clone", do_connection_clone, usage_connection_clone, TRUE, TRUE },
{ "import", do_connection_import, usage_connection_import, TRUE, TRUE },
{ "export", do_connection_export, usage_connection_export, TRUE, TRUE },
{ "monitor", do_connection_monitor, usage_connection_monitor, TRUE, TRUE },
{ NULL, do_connections_show, usage, TRUE, TRUE },
cli/connections: use nmc_do_cmd()
2016-07-06 13:34:07 +02:00
};
nmcli: add --order option for 'nmcli connection show' The option allows you to specify custom sorting order. Default order (when no --order is provided) corresponds to -o "active:name:path" Examples: nmcli con show -o name nmcli con show -o +name - sort connections by name alphabetically nmcli con show -o -name - sort connections by name alphabetically in reverse order mmcli con show -o active:name - sort connections first by active status, then by name mmcli con show -o -path - sort connections by D-Bus path in reverse order
2015-02-16 18:07:12 +01:00
cli: simplify do_connections() now that we don't wait for "connection-read" nm_remote_settings_list_connections() ensures the connections are loaded first. We check "help" commands before getting connections, because it is time intensive action.
2014-08-22 20:19:13 +02:00
/* Entry point function for connections-related commands: 'nmcli connection' */
NMCResultCode
do_connections (NmCli *nmc, int argc, char **argv)
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
{
nmcli: allow cmd specific option parsing in next_arg() func Options dependant on specific commands (e.g., nmcli connection show --active) are now allowed to be processed by the next_arg() function. This would allow autocompletion to expand options belonging to specific command first, and then global ones. Note that global options ("--ask" and "--show-secrets") will be auto-completed everywhere but only if at least a '-' is passed. Command specific ones (--temporary, --active, --order) will be auto-completed only after the command they belongs to but without requiring the user to pass a heading '-'. Example: 'nmcli connection show -a' will expand '-a' into '--active', but 'nmcli connection add -a` will expand '-a' into '--ask' (as it is a global option) This commit fixes also autocompletion for: nmcli connection modify --temporary
2017-03-30 16:09:46 +02:00
next_arg (nmc, &argc, &argv, NULL);
cli: add 'nmcli agent' command (bgo #739568) Synopsis: nmcli agent { secret | polkit | all } The command runs separate NetworkManager secret agent or session polkit agent, or both. It is useful when - no other secret agent is available (such as GUI nm-applet, gnome-shell, KDE applet) - no other polkit agent is available (such as polkit-gnome-authentication-agent-1, polkit-kde-authentication-agent-1 or lxpolkit) https://bugzilla.gnome.org/show_bug.cgi?id=739568
2014-10-30 15:45:43 +01:00
/* Register polkit agent */
nmc_start_polkit_agent_start_try (nmc);
cli: simplify do_connections() now that we don't wait for "connection-read" nm_remote_settings_list_connections() ensures the connections are loaded first. We check "help" commands before getting connections, because it is time intensive action.
2014-08-22 20:19:13 +02:00
/* Set completion function for 'nmcli con' */
cli: switch to the new-style readline typedefs (bgo #731294) The old Function, CPFunction, CPPFunction and VFunction typedefs are considered deprecated starting from readline 4.2 (though were not marked as such): http://lists.gnu.org/archive/html/bug-bash/2000-10/msg00078.html Compatibility typedefs have been in place up to readline 6.2 but were removed with the 6.3 release thus causing build to break. http://lists.gnu.org/archive/html/bug-readline/2014-04/msg00018.html https://bugzilla.gnome.org/show_bug.cgi?id=731294
2014-06-06 11:22:24 +02:00
rl_attempted_completion_function = (rl_completion_func_t *) nmcli_con_tab_completion;
cli: use readline also in other commands (besides interactive editor) It is used for '--ask' option. The main use is for questionnaires (nmcli -a con add).
2014-05-14 09:22:50 +02:00
cli: make it possible to call sub-commands with client obtained asynchronously
2016-08-31 21:04:33 +02:00
nmc_do_cmd (nmc, connection_cmds, *argv, argc, argv);
return nmc->return_value;
cli: add initial pieces of nmcli
2010-02-25 09:52:30 -08:00
}
cli: add nmcli monitor https://bugzilla.redhat.com/show_bug.cgi?id=1034158
2015-03-27 13:07:43 +01:00
void
monitor_connections (NmCli *nmc)
{
do_connection_monitor (nmc, 0, NULL);
}
Reference in a new issue Copy permalink