For tests based on glib's test framework, you can select which tests to
run by passing -p/-s on the command line.
Usually, we want to invoke our tests via `make check` which conveniently
calls valgrind (run-test-valgrind.sh) or spawns a private D-Bus server
(libnm-test-launch.sh, libnm-glib-test-launch.sh). At that point, it is
not directly possible to specify command line arguments for the tests,
which is why it is convenient to specify arguments via $NMTST_DEBUG
environment variable.
Parse "p" and "s" arguments from $NMTST_DEBUG and pass them to g_test_init()
to select which tests to run.
NMTST_DEBUG=p=/core/general/test_setting_ip4_changed_signal ./libnm-core/tests/test-general
However, there is a problem here: in gtestutils, -p/-s conflicts with --tap,
which is how our Makefile invokes the tests. Thus the new options explicitly
don't work when being called during `make check`. Which makes this much
less useful. I only noticed that afterwards, so still keep the patch
because it might still be convenient during developing tests to set the
environment variable once, and then repeatedly spawn the tests, without
specifying -p/-s.
(cherry picked from commit 73cb579108)
If __x is signed and its value is the minimum allowed for the type,
(__x - 1) causes a signed overflow, which has an undefined
behavior. Compiling with -fsanitize=undefined produces the warnings:
test-general.c:4499:619: runtime error: signed integer overflow: -2147483648 - 1 cannot be represented in type 'int'
test-general.c:4506:681: runtime error: signed integer overflow: -2147483648 - 1 cannot be represented in type 'TEST_IS_POWER_OF_TWP_ENUM_SIGNED'
test-general.c:4501:619: runtime error: signed integer overflow: -9223372036854775808 - 1 cannot be represented in type 'long int'
test-general.c:4509:691: runtime error: signed integer overflow: -9223372036854775808 - 1 cannot be represented in type 'TEST_IS_POWER_OF_TWP_ENUM_UNSIGNED_64'
Change the order of expressions to avoid this.
Fixes: 0a3c1f5774
(cherry picked from commit bc9d034775)
It is valid to call nm_clear_g_signal_handler() with missing
@self argument if (and only if) the @id is unspecified as well.
Remove the check for @self to get an assertion in case @id
is missing *and* @self is invalid. In this case,
g_signal_handler_disconnect() will raise a g_critical() for us.
Fixes: c33416178f
(cherry picked from commit 76958f94d4)
It does not make sense to issue an error. This should be a helper function.
"NetworkManager[18341]: nm_clear_g_signal_handler: assertion 'G_IS_OBJECT (self)' failed"
error started since commit e6d7fee5a6 due to that.
(cherry picked from commit c33416178f)
For libnm library, "nm-dbus-interface.h" contains defines like the D-Bus
paths of NetworkManager. It is desirable to have this header usable without
having a dependency on "glib.h", for example for a QT application. For that,
commit c0852964a8 removed that dependancy.
For libnm-glib library, the analog to "nm-dbus-interface.h" is
"NetworkManager.h", and the same applies there. Commit
159e827a72 removed that include.
However, that broke build on PackageKit [1] which expected to get the
version macros by including "NetworkManager.h". So at least for libnm-glib,
we need to preserve old behavior so that a user including
"NetworkManager.h" gets the version macros, but not "glib.h".
Extract the version macros to a new header file "nm-version-macros.h".
This header doesn't include "glib.h" and can be included from
"NetworkManager.h". This gives as previous behavior and a glib-free
include.
For libnm we still don't include "nm-version-macros.h" to "nm-dbus-interface.h".
Very few users will actually need the version macros, but not using
libnm.
Users that use libnm, should just include (libnm's) "NetworkManager.h" to
get all headers.
As a special case, a user who doesn't want to use glib/libnm, but still
needs both "nm-dbus-interface.h" and "nm-version-macros.h", can include
them both separately.
[1] https://github.com/hughsie/PackageKit/issues/85
Fixes: 4545a7fe96
(cherry picked from commit 7bf10a75db)
All current users of NM_IN_SET() would rather use short-circuit evalation
(or don't care). It seems that doing it by default seems favorable.
The only downside is, that this might have somewhat unexpected behavior
to a user who expects a regular function (which would evaluate always
all arguments).
Fixes: 7860ef959a
(cherry picked from commit 96cacc07e8)
Let the preprocessor do more work, but generate a simple expression that
the compiler can optimize (presumably) better.
(cherry picked from commit 7860ef959a)
Code that is testable often needs special hooks to work
both for unit-tests and production.
Add a function nm_utils_get_testing() that returns whether
the code is run as part of a unit-test.
For non-testing mode, nm_utils_get_testing() will return
zero (NM_UTILS_TEST_NONE). For unit tests, the test should call
_nm_utils_set_testing() to configure tested functions.
By specifing the @flags attribute, the test can enable/disable
specific behaviors.
https://bugzilla.gnome.org/show_bug.cgi?id=701112
(cherry picked from commit b9d8dc050a)
We already have "nm-utils*.h" and "NetworkManagerUtils.h" headers. Rename
"include/nm-utils-internal.h" to "nm-macros-internal.h". I think that
name is better, because this file is header-only, internal, and
repository-wide.
Also, it will never contain non-header-only declarations because
there is no backing object file under "include/".
It will only contain macros and inline functions.
(cherry picked from commit b8b1a01d96)
Even if asserts are not enabled, still let the compiler see what we would
assert. Otherwise, we get warnings about unused variables or we migth miss
compile errors inside nm_assert().
(cherry picked from commit fe3e1849b7)
NM already has two kinds of assertions:
- g_assert*(), conditionally compiled via #ifndef G_DISABLE_ASSERT
- g_return*(), conditionally compiled via #ifndef G_DISABLE_CHECKS
In theory, one should be able to disable both asserts and NM should
still work correctly (and possibly more efficient). In practice,
hardly anybody is testing such a configuration and it might be broken.
Especially, we don't disable asserts for production builds, both because
of less test coverage and because it might reduce our ability to debug.
Add a new configure option --enable-more-asserts, which defines
NM_MORE_ASSERTS and nm_assert(). This is for expensive asserts,
that -- contrary to the asserts above -- are disabled by default.
This is useful for extended debugging.
(cherry picked from commit 08ecafd2bf)
And nmtst_platform_ip6_routes_equal().
As already indicated by the comment, using memcmp() doesn't really work
here. It worked up to now, because the NMPlatformIP4Route structure has no
padding at the end.
If we would have a last 'guint8' member (as we will add later),
initializing an arry of routes on the stack would not clear the bytes
from the padding and the comparison using memcmp() would fail.
(cherry picked from commit d8d0c481b8)
glib interprets the options either as "-m arg" or "-m=arg". Fix parsing
to check for both cases.
Also, g_test_init() removes the parsed options from argv, hence we must
check our original copy in __nmtst_internal.orig_argv.
Now the following all have the same outcome:
$ NMTST_DEBUG=no-debug,quick ./src/rdisc/tests/test-rdisc-fake
$ ./src/rdisc/tests/test-rdisc-fake -m quick
$ ./src/rdisc/tests/test-rdisc-fake -m=quick
Fixes: a2edd6445f
(cherry picked from commit f529fedd17)
Utility function for long-running tests to check whether
they should run. It has the same use as g_test_quick(),
but it can also be set via NMTST_DEBUG environment variable.
You can set the compile time default via
CFLAGS='-DNMTST_TEST_QUICK=FALSE' ./configure
and overwrite it at runtime via
NMTST_DEBUG=no-default,quick ./tests/test-my
Utility function to simplify the following common code:
if (priv->timeout_id) {
g_source_remove (priv->timeout_id);
priv->timeout_id = 0;
}
to
nm_clear_g_source (&priv->timeout_id);
(cherry picked from commit b5beaef8fa)
For tests with assert-logging enabled, we already printed
the message with g_print(), instead of g_message().
We always want to print this value, because it is important
to reproduce a test failure. Hence, just (always) print
directly to stdout.
Also, add a leading newline, because the output was attached
to the previous line.
(cherry picked from commit 4ab540939c)
When disabling assert-logging with no-expect-message,
print a line at every g_test_assert_expected_messages()
invocation.
(cherry picked from commit 373d09b042)
On the end of the test there should be no unchecked expected
messages. Register a destructor function in nmtst to assert
against that.
Also call nmtst_free() in the destructor. That way, we always release
the allocated data without requiring the test to do it explicitly.
(cherry picked from commit bb90127c2a)
Same for nmtst_platform_ip6_routes_equal().
It's useful to check for equal routes ignoring the ordering.
(cherry picked from commit 57453189e0)
Conflicts:
src/tests/test-route-manager.c
Before, when having a test with nmtst_init_assert_logging(),
the caller was expected to setup logging separately according
to the log level that the test asserts against.
Since 5e74891b58, the logging
level can be reset via NMTST_DEBUG also for tests that
assert logging. In this case, it would be useful, if the test
would not overwrite the logging level that is set externally
via NMTST_DEBUG.
Instead, let the test pass the logging configuration to
nmtst_init_assert_logging(), and nmtst will setup logging
-- either according to NMTST_DEBUG or as passed in.
This way, setting the log level works also for no-expect-message
tests:
NMTST_DEBUG="debug,no-expect-message,log-level=TRACE" $TEST
(cherry picked from commit b6d3b98655)
For glog messages to print any debug messages, we must set G_MESSAGES_DEBUG.
nmtst does this for us if we set @is_debug. But fix the condition to
also set G_MESSAGES_DEBUG if set set c_log_level to DEBUG or TRACE.
(cherry picked from commit 3abeb7f2b9)
