Bug 378126 – gimp crash due to GTK

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 378126 - gimp crash due to GTK


Summary:	gimp crash due to GTK


Status:	RESOLVED DUPLICATE of bug 363472

Product:	gtk+
Classification:	Platform
Component:	Widget: GtkFileChooser
Version:	2.10.x
Hardware:	Other Linux

Importance:	Normal critical
Target Milestone:	---
Assigned To:	gtk-bugs
QA Contact:	Federico Mena Quintero

URL:
Whiteboard:

Depends on:
Blocks:

Reported:	2006-11-22 13:58 UTC by Sebastien Bacher
Modified:	2006-11-23 16:20 UTC

See Also:
GNOME target:	---
GNOME version:	2.15/2.16

Description Sebastien Bacher 2006-11-22 13:58:42 UTC

That bug has been described on https://launchpad.net/distros/ubuntu/+source/gimp/+bug/72633

"I was working with gimp , and than just switched to work on a different desktop.
Suddenly, the gimp crashed - I didn't touched it.

http://librarian.launchpad.net/5116345/_usr_bin_gimp.1000.crash
crash
...
 libgtk2.0-0 2.10.6-0ubuntu1
..."


Debug backtrace:

(gdb) bt

+ Trace 88504

#0 IA__gtk_file_info_get_is_folder
at gtkfilesystem.c line 177
#1 name_sort_func
at gtkfilechooserdefault.c line 5598
#2 gtk_tree_model_sort_level_find_insert
at gtktreemodelsort.c line 1739
#3 gtk_tree_model_sort_row_inserted
at gtktreemodelsort.c line 1787
#4 _gtk_marshal_VOID__BOXED_BOXED
at gtkmarshalers.c line 1305
#5 IA__g_closure_invoke
at gclosure.c line 490
#6 signal_emit_unlocked_R
at gsignal.c line 2440
#7 IA__g_signal_emit_valist
at gsignal.c line 2199
#8 IA__g_signal_emit
at gsignal.c line 2243
#9 IA__gtk_tree_model_row_inserted
at gtktreemodel.c line 1494
#10 do_files_added
at gtkfilesystemmodel.c line 1755
#11 IA__g_cclosure_marshal_VOID__POINTER
at gmarshal.c line 601
#12 IA__g_closure_invoke
at gclosure.c line 490
#13 signal_emit_unlocked_R
at gsignal.c line 2440
#14 IA__g_signal_emit_valist
at gsignal.c line 2199
#15 IA__g_signal_emit_by_name
at gsignal.c line 2267
#16 monitor_callback
at gtkfilesystemgnomevfs.c line 3464
#17 actually_dispatch_callback
at gnome-vfs-monitor.c line 271
#18 g_idle_dispatch
at gmain.c line 3926
#19 IA__g_main_context_dispatch
at gmain.c line 2045

Comment 1 Sebastien Bacher 2006-11-22 14:04:27 UTC

From gdb:

(gdb) p info
$3 = (const GtkFileInfo *) 0x2f726176
(gdb) p info->is_folder
Cannot access memory at address 0x2f726176

Comment 2 Frank Arnold 2006-11-22 14:10:16 UTC

Duplicate of bug 363472

Comment 3 Sebastien Bacher 2006-11-22 17:59:33 UTC

why do you think that's a duplicate? it's listed by the dupfinder, the backtraces look different though

Comment 4 Frank Arnold 2006-11-22 18:58:41 UTC

(In reply to comment #3)
> why do you think that's a duplicate? it's listed by the dupfinder, the
> backtraces look different though
> 

Because GTKFileInfo always has the state like you described in comment #1. On one of those crashes I did see it as a *ghost* entry with no name in the filechooser. It then crashed by clicking on it. So I assume the pointer shouldn't exist anymore. Traces are slightly different, but the root cause should be the same.

Anyways, the one I fetched with debuginfo in [1] looks very similar to your backtrace. Before installing debuginfo packages it looked like the original one of bug 363472

But feel free to keep it open. I don't have much knowledge about C and GTK.

[1] http://bugzilla.gnome.org/attachment.cgi?id=76832&action=view

Comment 5 Sebastien Bacher 2006-11-23 16:20:34 UTC

explanation makes sense, I'm marking it as duplicate, thank you for the comment

*** This bug has been marked as a duplicate of 363472 ***