This fixes a race condition: the server exits while the client continues
to the next iteration. If the server wins, the test passes. If the client
wins, it sends a message to the dying service, never gets a reply, and the
test fails.
My branch to refactor the main loop for fd.o #23194 seems to make the
client more likely to win this race, resulting in intermittent test
failures.
This is an instance of the general problem described by fd.o #11454.
This is similar to how ConnectionData works. Without this change, we
deserve to segfault: when the first set of callbacks (either watches or
timeouts) is cleaned up, we unref the server and loop, and free sd;
when the second set of callbacks is cleaned up, we use-after-free sd,
the server and the loop, then double-free sd.
However, due to fd.o #33277 we don't even get that far, because we've
already died with an assertion failure.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33277
Helpfully, AC_CHECK_DECLS (which configure.in uses to define
HAVE_DECL_MSG_NOSIGNAL) behaves differently to every similar AC_CHECK_*
macro and, rather than producing #undef HAVE_DECL_MSG_NOSIGNAL if the
given symbol is not found, instead produces #define
HAVE_DECL_MSG_NOSIGNAL 0.
Mike McQuaid's patch fixes the uses of this constant in sysdeps-unix;
with this patch, all code is guarded consistently and correctly.
Add DBUS_INVALID_NESTED_TOO_DEEPLY validity problem and a test that
should generate it.
Previously, we rejected deep nesting in the signature, but
variants allow dynamic message nesting, conditional only
on the depth of the message body.
The nesting limit is 64, which was also the limit in static
signatures. Empirically, dynamic nesting depth observed on my
Fedora 14 system doesn't exceed 2; 64 is really a huge limit.
https://bugs.freedesktop.org/show_bug.cgi?id=32321
Signed-Off-By: Colin Walters <walters@verbum.org>
Signed-off-by: Will Thompson <will.thompson@collabora.co.uk>
MSG_NOSIGNAL could theoretically be an enum member or something rather
than an #define, so it's better to check for the flag defined by the
configure check (as is done in dbus-connection.c already).
Reviewed-by: Colin Walters <walters@verbum.org>
Reviewed-by: Scott James Remnant <scott@netsplit.com>
Since SIGPIPE is no longer touched by default when MSG_NOSIGNAL is
available, it's extra-critical that all socket writes actually pass
that flag.
Signed-off-by: Will Thompson <will.thompson@collabora.co.uk>
This reverts commit 4626b40560.
test-autolaunch works fine in non-launchd environments (and non-X11
environments, based on a quick test passing enable_x11=no to configure).
On the contrary: this commit *broke* the build on non-launchd
environments, because test/name-test/run-test.sh still tried to run this
test even if it hadn't been built.
The excellently-titled commit 197bef8 “Fix test failures on OSX.” broke
the tests on Linux, since there's no wheel group on this side of the
tracks. So here's a group everyone should enjoy.
(If anyone comes along and tells me that DragonflyBSD doesn't have
'nogroup' …)
