callisto-debian/evolution
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Instead of doing pthread_exit() after redirecting the SEGV, try to lock a

Browse Source 
* main.c (segv_redirect): Instead of doing pthread_exit() after
	redirecting the SEGV, try to lock a mutex we know is already
	locked. This will hopefully help debug a bunch of bugs where the
	crashed thread seems to be missing from the bug-buddy report.

svn path=/trunk/; revision=14699
This commit is contained in:
Dan Winship
2001-11-14 20:26:00 +00:00
parent 4647e0c39c
commit 417af81526
2 changed files with 17 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
mail/ChangeLog
View File

@ -1,3 +1,10 @@
2001-11-14 Dan Winship <danw@ximian.com>
* main.c (segv_redirect): Instead of doing pthread_exit() after
redirecting the SEGV, try to lock a mutex we know is already
locked. This will hopefully help debug a bunch of bugs where the
crashed thread seems to be missing from the bug-buddy report.
2001-11-13 Jeffrey Stedfast <fejj@ximian.com>
* mail-session.c (pass_got): Call e_passwords_remember_password()

11
mail/main.c
View File

@ -62,6 +62,8 @@ static int blowup(int status)
*/
static void (*gnome_segv_handler) (int);
static GStaticMutex segv_mutex = G_STATIC_MUTEX_INIT;
static void
segv_redirect (int sig)
{
@ -69,7 +71,13 @@ segv_redirect (int sig)
gnome_segv_handler (sig);
else {
pthread_kill (mail_gui_thread, sig);
pthread_exit (NULL);
/* We can't return from the signal handler or the
* thread may SEGV again. But we can't pthread_exit,
* because then the thread may get cleaned up before
* bug-buddy can get a stack trace. So we block by
* trying to lock a mutex we know is already locked.
*/
g_static_mutex_lock (&segv_mutex);
}
}
@ -105,6 +113,7 @@ main (int argc, char *argv [])
sigaction (SIGBUS, &sa, NULL);
sigaction (SIGFPE, &sa, NULL);
gnome_segv_handler = osa.sa_handler;
g_static_mutex_lock (&segv_mutex);
}
orb = oaf_init (argc, argv);