On Thu, Mar 10, 2016 at 04:29:40PM +0000, Frediano Ziglio wrote:
> Fatal error exit usually with abort() causing SIGABRT to be triggered
> which can have problems with core dump generation depending on system
> settings. Capturing and exiting on this signal solve the problem.
I still think this looks like a workaround for an issue elsewhere in the
stack. If the goal is to prevent core dump generation, this deserves a
comment in the code near handle_sigabrt as doing this is very unusual
and seems redundant with what is going to happen when SIGABRT is raised.
Christophe
>
> Signed-off-by: Frediano Ziglio <fziglio@xxxxxxxxxx>
> ---
> tests/test-logging.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/tests/test-logging.c b/tests/test-logging.c
> index 94e923a..72f35a4 100644
> --- a/tests/test-logging.c
> +++ b/tests/test-logging.c
> @@ -377,11 +377,17 @@ static void test_spice_g_messages_debug_all(void)
> g_test_trap_assert_stderr("*g_message\n*other_message\n");
> }
>
> +static void handle_sigabrt(int sig)
> +{
> + _Exit(1);
> +}
>
> int main(int argc, char **argv)
> {
> GLogLevelFlags fatal_mask;
>
> + signal(SIGABRT, handle_sigabrt);
> +
> fatal_mask = (GLogLevelFlags)g_log_set_always_fatal((GLogLevelFlags) G_LOG_FATAL_MASK);
>
> g_test_init(&argc, &argv, NULL);
> --
> 2.5.0
>
> _______________________________________________
> Spice-devel mailing list
> Spice-devel@xxxxxxxxxxxxxxxxxxxxx
> https://lists.freedesktop.org/mailman/listinfo/spice-devel
Attachment:
signature.asc
Description: PGP signature
_______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel