NetworkManager/src/core/settings/plugins
Thomas Haller f15498eda3
ifcfg-rh: cleanup make_bond_port_setting()
- use svGetValue() instead of svGetValueStr(). The difference is that
  svGetValueStr() coerces "" to NULL. "" is not a valid value, but we
  want to parse the value and print an warning message about it. Also,
  the presence of the variable determines whether we add the bond-port
  setting or not.

- don't use nm_clear_g_free(). @value_to_free is gs_free, it will be
  cleared automatically.

- use g_object_set() instead of nm_g_object_set_property_uint(). The
  latter is our own implementation that does error checking (e.g., that
  the value is in range (0..2^16-1). But we already ensured that to
  be the case. So just call g_object_set(), it cannot fail and if it
  would, we want the assertion failure that it would cause.

- queue_id should be a "guint". It is always true on Linux/glib that
  sizeof(guint) >= sizeof(guint32), the opposite theoretically might not
  be true.
  But later we use the variable in the variadic function g_object_set(),
  where it should be guint.

- the errno from _nm_utils_ascii_str_to_uint64() isn't very useful for
  logging. It's either ERANGE or EINVAL, and logging the numeric values
  of these error codes isn't gonna help the user. We could stringify
  with nm_strerror_native(errno), but that message is also not very
  useful. Just say that the string is not a number.
2021-08-26 23:05:17 +02:00
..
ifcfg-rh ifcfg-rh: cleanup make_bond_port_setting() 2021-08-26 23:05:17 +02:00
ifupdown all: rename nm_utils_strdict_*() to nm_strdict_*() 2021-08-02 09:26:48 +02:00
keyfile all: rename nm_utils_strbuf_*() API to nm_strbuf_*() 2021-08-02 09:26:42 +02:00
meson.build all: move "src/" directory to "src/core/" 2021-02-08 09:56:41 +01:00