Bug 708647 – Totem segfaults on start

After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.

Bug 708647 - Totem segfaults on start


Summary:	Totem segfaults on start


Status:	RESOLVED FIXED

Product:	totem
Classification:	Core
Component:	general
Version:	3.10.x
Hardware:	Other Linux

Importance:	Normal critical
Target Milestone:	---
Assigned To:	General Totem maintainer(s)
QA Contact:	General Totem maintainer(s)

URL:
Whiteboard:

Duplicates:	708747 (view as bug list)
Depends on:
Blocks:

Reported:	2013-09-23 18:01 UTC by Dominique Leuenberger
Modified:	2013-09-27 13:58 UTC

See Also:
GNOME target:	---
GNOME version:	---

Attachments
A valgrind trace / compressed, as it's 3.8MB (72.23 KB, application/x-xz) 2013-09-23 21:36 UTC, Dominique Leuenberger	Details

Description Dominique Leuenberger 2013-09-23 18:01:28 UTC

Totem segfaults on startup...

(gdb) bt

+ Trace 232530

#0 __GI_raise
at ../nptl/sysdeps/unix/sysv/linux/raise.c line 56
#1 __GI_abort
at abort.c line 89
#2 _g_log_abort
at gmessages.c line 255
#3 g_assertion_message
#4 g_assertion_message_expr
at gtestutils.c line 2293
#5 totem_fullscreen_show_popups
at totem-fullscreen.c line 310
#6 totem_object_action_remote
at totem-object.c line 3122
#7 totem_options_process_for_server
at totem-options.c line 230
#8 app_command_line
at totem.c line 246
#9 ffi_call_unix64
at ../../../libffi/src/x86/unix64.S line 75
#10 ffi_call
at ../../../libffi/src/x86/ffi64.c line 492
#11 g_cclosure_marshal_generic
at gclosure.c line 1454
#12 g_closure_invoke
at gclosure.c line 777
#13 signal_emit_unlocked_R
at gsignal.c line 3586
#14 g_signal_emit_valist
at gsignal.c line 3340
#15 g_signal_emit
at gsignal.c line 3386
#16 g_application_run
at gapplication.c line 1612
#17 main
at totem.c line 291




Starting it with "totem --fullscreen" avoids the crash; but does not start in fullscreen.

Comment 1 Dominique Leuenberger 2013-09-23 18:04:33 UTC

starting it on CLI shows:

> totem

(totem:9989): Gtk-CRITICAL **: gtk_action_set_icon_name: assertion 'GTK_IS_ACTION (action)' failed
**
Totem:ERROR:totem-fullscreen.c:310:totem_fullscreen_show_popups: assertion failed: (fs->priv->is_fullscreen != FALSE)
Aborted (core dumped)

Comment 2 Bastien Nocera 2013-09-23 18:30:49 UTC

I've fixed the warning, but the crasher makes no sense. How do I reproduce it exactly?

commit 437e64ffb5f19d2f088339954686e794cc276319
Author: Bastien Nocera <hadess@hadess.net>
Date:   Mon Sep 23 20:27:17 2013 +0200

    main: Fix typo in action name
    
    See: https://bugzilla.gnome.org/show_bug.cgi?id=708647

Comment 3 Dominique Leuenberger 2013-09-23 18:38:39 UTC

(In reply to comment #2)
> I've fixed the warning, but the crasher makes no sense. How do I reproduce it
> exactly?

I wish I knew :(
totem worked fine for me forever.. and today, after I updated my packages (openSUSE) to 3.10.0, I see crashes (downgrading does not change it)

it's the regular way of starting it...

Comment 4 Bastien Nocera 2013-09-23 18:50:21 UTC

You'll need to run this under valgrind, because this is impossible:

+ Trace 232531

#6 totem_object_action_remote
at totem-object.c line 3122


cmd has a max value of 28, as it's a TotemRemoteCommand enum.

Comment 5 Dominique Leuenberger 2013-09-23 20:16:14 UTC

I found a way to reproduce it (and as such, also a way to avoid it);
* Enable the Zeitgeist plugin (which newly re-appeared in 3.9.92 / 3.10) and the crash happens.

After disabling the plugin, totem starts up fine again.

Comment 6 Dominique Leuenberger 2013-09-23 20:27:19 UTC

Program received signal SIGABRT, Aborted.
0x00007ffff5dba849 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
56	../nptl/sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) thread apply all bt

+ Trace 232532

Thread 1 (Thread 0x7ffff7f9aa00 (LWP 25095))

#0 __GI_raise
at ../nptl/sysdeps/unix/sysv/linux/raise.c line 56
#1 __GI_abort
at abort.c line 89
#2 _g_log_abort
at gmessages.c line 255
#3 g_assertion_message
#4 g_assertion_message_expr
at gtestutils.c line 2293
#5 totem_fullscreen_show_popups
at totem-fullscreen.c line 310
#6 totem_object_action_remote
at totem-object.c line 3122
#7 totem_options_process_for_server
at totem-options.c line 230
#8 app_command_line
at totem.c line 246
#9 ffi_call_unix64
at ../../../libffi/src/x86/unix64.S line 75
#10 ffi_call
at ../../../libffi/src/x86/ffi64.c line 492
#11 g_cclosure_marshal_generic
at gclosure.c line 1454
#12 g_closure_invoke
at gclosure.c line 777
#13 signal_emit_unlocked_R
at gsignal.c line 3586
#14 g_signal_emit_valist
at gsignal.c line 3340
#15 g_signal_emit
at gsignal.c line 3386
#16 g_application_run
at gapplication.c line 1612
#17 main
at totem.c line 291

Comment 7 Dominique Leuenberger 2013-09-23 21:36:48 UTC

Created attachment 255596 [details]
A valgrind trace / compressed, as it's 3.8MB

Comment 8 Bastien Nocera 2013-09-25 13:56:40 UTC

*** Bug 708747 has been marked as a duplicate of this bug. ***

Comment 9 Bastien Nocera 2013-09-27 13:54:14 UTC

Had nothing to do with zeitgeist but with whether the window was
focused by the time we handled command-line options. Fixed in gnome-3-8 and gnome-3-10.

Comment 10 Dominique Leuenberger 2013-09-27 13:58:21 UTC

(In reply to comment #9)
> Had nothing to do with zeitgeist but with whether the window was
> focused by the time we handled command-line options. Fixed in gnome-3-8 and
> gnome-3-10.

Nice catch... so likely 'more plugins loaded' increased the chances of not being focused 'yet'; exposing the bug.

THANKS for the fix!