GNOME Bugzilla – Bug 639717
Crash imapx_command_select_done at camel-imapx-server.c:2490
Last modified: 2013-08-22 17:06:49 UTC
evolution 2.91.5 glib master gtk 2.24 1. Created a filter 2. Pressed ctrl+y 3. Status bar says 'filtering...' 4. Waited for more than 10 minutes but it didn't filter around 150 mails 5. Pressed Cancel on tool bar and evolution crashed. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0xab3ffb70 (LWP 8817)] 0xad3e3263 in imapx_command_select_done (is=0x8573010, ic=0xa9304798) at camel-imapx-server.c:2490 2490 ic->status->text? ic->status->text:"<unknown reason>"); (gdb) t a a bt
+ Trace 225560
Thread 1 (Thread 0xb5fe8830 (LWP 8801))
(gdb)
Downstream bug report about the same in 2.32.1: https://bugzilla.redhat.com/show_bug.cgi?id=676391
we also have a report about this here: https://bugs.launchpad.net/ubuntu/+source/evolution/+bug/767086
evolution 3.1.2 Program received signal SIGSEGV, Segmentation fault.
+ Trace 227357
Thread 2977512304 (LWP 2758)
I think I've got hit by this problem too. Running Evo 3.1.4 with e-d-s 3.1.5 I just got a crash with the following backtrace: Program received signal SIGSEGV, Segmentation fault.
+ Trace 228227
Thread 140735626057472 (LWP 21730)
It tries to access ic->status->text but ic->status is NULL: (gdb) print ic->status $20 = (struct _status_info *) 0x0 I guess we should protect that path if with if (ic->status) or figure why status is 0x0 in the first place.
Created attachment 194802 [details] [review] Patch to deal with ic->status being 0x0 The attached patch makes sure we can cope with ic->status being NULL (which is what is triggering the crash).
Patch can be used as a workaround till proper fix, and why it happened, will be found. Only add a g_debug() when the ic->status is NULL, to know that it happened, with a FIXME comment above it with a little bit explanation.
Merged the workaround: commit 094ab8f70b5b9750448c8b65ca32d5282c886460 Author: Raul Gutierrez Segales <rgs@collabora.co.uk> Date: Fri Aug 26 10:36:02 2011 +0100 Check ic->status is not NULL imapx_command_select_done () There are ocassions (are they even valid?) in which ic->status might be NULL, so we check for that before trying to access ic->status->text when calling g_set_error (). Fixes: https://bugzilla.gnome.org/show_bug.cgi?id=639717 Should we close this or leave it open until we find the root cause?
*** Bug 657847 has been marked as a duplicate of this bug. ***
last dup in 3.1.5.x
Closing as OBSOLETE since the stack trace(s) are too old to be useful at this point.