Created attachment 90543 [details]
Occasionally, sending SIGINT to weston results in weston crashing with SIGSEGV during shell_destroy listener execution. See attached backtrace. I was able to trigger this about 50% of the time.
wayland (master) heads/master-0-g1521c62
drm (master) libdrm-2.4.50-0-g4c5de72
mesa (master) heads/master-0-g0319ea9
libva (master) heads/master-0-g73a11b3
intel-driver (master) heads/master-0-g9d0bd94
weston (master) heads/master-0-g8d344a0
Also observed this on the Weston 1.3.1 stable stack:
wayland (1.3) heads/1.3-0-g84247b7
drm (master) libdrm-2.4.49-0-g7ddc98f
mesa (10.0) heads/10.0-0-gb160fea
libva (master) libva-1.2.1-0-g88ed1eb
intel-driver (master) 1.2.1-0-g8f306e3
weston (1.3) 1.3.1-0-g5e3ff49
Author: Ander Conselvan de Oliveira <firstname.lastname@example.org>
Date: Fri Dec 20 21:07:01 2013 +0200
shell: Handle the desktop shell client destroy signal
Set the internal pointer for the client to NULL. This fixes a
segmentation fault at shutdown, where the shell would hang up before
and cause libwayland to call wl_client_destroy(). When the shell was
destroyed later, another call to wl_client_destroy() would cause the
Verified fixed on weston (master) heads/master-0-g917711b...
...will this be backported to 1.3?