There's a race here, as revealed by Debian's buildds.
We call g_dbus_proxy_new() to create a proxy for the test server, with
callback proxy_ready() Then we call g_spawn_command_line_async() to
start the test server, and then start the main loop.
proxy_ready() assumes that the test server hasn't been started when it
is called. But there is no guarantee that these asynchronous operations
involving spawning a process won't happen in a different order that mean
the bus name *does* have an owner.
What we can do is move starting the server inside of proxy_ready(), so
we know that the test server isn't started until after the proxy is
created. We also add an assertion to check that it is indeed not running
before we execute it.
The subprocess needs to access the test_log_fd. If the file descriptors
are not left open, functions such as g_test_message may stomp on file
descriptors open by the subprocess and cause bad behavior of the test.
(Tweaked by Philip Withnall <bugzilla@tecnocode.co.uk> to fix review
comments.)
In C++ we can use nullptr to ensure g_assert_[non]null() is only called
with pointers. This will introduce build failures in tests that would
have previously compiled, but only in C++, and only for code that
misused these macros. Code using the macros properly will be fine.
This change caught a couple bugs in WebKit's API tests, where I had
accidentally used these functions improperly. E.g. this is now a build
failure in C++:
g_assert_null(webkit_context_menu_get_n_items(menu)); /* Oops! */
Either I wanted to use cmpuint there, or I wanted to use
webkit_context_menu_get_items() to receive a GList* instead.
Another example that will no longer build in C++:
g_assert_null(0); /* Contrived, but 0 is not a pointer! */
So long, and thanks for everything. We’re a Meson-only shop now.
glib-2-58 will remain the last stable GLib release series which is
buildable using autotools.
We continue to install autoconf macros for autotools-using projects
which depend on GLib; they are stable API.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
The behaviour of the Meson build has changed a little vs what we did in
autotools. In autotools, --enable-debug was a tristate (yes, no,
undefined), with all three options resulting in different macro
definitions.
In Meson, we have a bistate of --buildtype={debug,debugoptimized} vs
--buildtype=(anything else). There is no way to automatically define
G_DISABLE_ASSERT or G_DISABLE_CHECKS while building GLib — you need to
define them in your CPPFLAGS in your environment instead.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
Drop mentions of autotools. In particular, update the list of configure
options to reflect what’s available in the Meson build.
Further work is needed as a follow-up to improve our handling of (what
was formerly) the --enable-debug option.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
We’re about to drop autotools support. Rather than keep the .mk files
around in master indefinitely, link to the versions in the glib-2-58
branch (the last stable release of GLib which supports building with
autotools) in readiness for dropping the .mk files from master.
Any future fixes to these files can happen on the glib-2-58 branch. The
links should work forever (as long as we use GitLab).
Signed-off-by: Philip Withnall <withnall@endlessm.com>
We don’t actually build this; the Makefile was just there to allow
ad-hoc regeneration of the glib-mirroring-tab output files.
Port it to Meson just so there are no remnants of GNU make left in GLib.
Don’t hook it up to the rest of the build.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
But it can't be used as a drop-in implementation of G_GNUC_NORETURN
because it can only be placed at the start of the function prototype.
Document this in a comment so that the next person doesn't spend
20 min figuring it out.
This effectively renders those tests useless (since realistically nobody
runs tests locally), but it’s better than every other CI run failing for
unrelated reasons. The idea is that the ‘flaky’ tag can be temporarily
applied to a test while a problem is being investigated or fixed, and
then removed later.
Signed-off-by: Philip Withnall <withnall@endlessm.com>
This is a wrapper around g_private_set() which allocates the desired
amount of memory for the caller and calls g_private_set() on it.
This is intended to make it easier to suppress Valgrind warnings about
leaked memory, since g_private_set() is typically used to make one-time
per-thread allocations. We can now just add a blanket suppression rule
for any allocations inside g_private_set_alloc0().
Signed-off-by: Philip Withnall <withnall@endlessm.com>
