GNOME Bugzilla – Bug 316018
Crash when printing gif file from gthumb
Last modified: 2005-09-21 17:32:07 UTC
Steps to reproduce:
1. Start gthumb
2. Open file mentionned at bug 316011
jgilli@susu:~/projets/bug_days/libgnomeprint$ gdb gthumb
GNU gdb 6.3-debian
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show warranty" for details.
This GDB was configured as "i386-linux"...Using host libthread_db library
Starting program: /home/jgilli/custom/bin/gthumb
[Thread debugging using libthread_db enabled]
[New Thread 16384 (LWP 16769)]
[New Thread 32769 (LWP 16770)]
[New Thread 16386 (LWP 16771)]
[New Thread 32771 (LWP 16772)]
[New Thread 49156 (LWP 16773)]
[New Thread 65541 (LWP 16774)]
[New Thread 81926 (LWP 16775)]
[New Thread 98311 (LWP 16776)]
[New Thread 114696 (LWP 16777)]
[New Thread 131081 (LWP 16778)]
[New Thread 147466 (LWP 16781)]
[New Thread 163851 (LWP 16782)]
(gthumb:16769): GnomePrintCupsPlugin-WARNING **: iconv does not support ppd
character encoding: ISOLatin1, trying CSISOLatin1
[Thread 147466 (LWP 16781) exited]
(gthumb:16769): GLib-GObject-WARNING **: IA__g_object_new_valist: object class
`GnomePrintDialog' has no property named `config'
(gthumb:16769): GLib-GObject-WARNING **: invalid (NULL) pointer instance
(gthumb:16769): GLib-GObject-CRITICAL **: g_signal_connect_data: assertion
`G_TYPE_CHECK_INSTANCE (instance)' failed
[New Thread 180236 (LWP 16783)]
[Thread 180236 (LWP 16783) exited]
Program received signal SIGSEGV, Segmentation fault.
I was trying to reproduce bug 316011.
Created attachment 52099 [details] [review]
It doesn't crash anymore
Although this patch prevent libgnomeprint from crashing, it doesn't print
Created attachment 52104 [details] [review]
Another try to prevent libgnomeprint from crashing
This one makes more sense.
Both print preview and print crash without the above patch on my workstation.
However, it seems that (from the cvs log output) this file hasn't changed since
08/15/2005, and many people are able to print from gthumb. I don't understand
how it is possible.
Moreover, as pointed out by oGALAXYo on #bugs, the above patch makes more sense.
!@$! that file has not changed in recent years! The code used to live in a
The patch does not look correct. I suspect that this was a cut-n-paste error.
It looks like the author was trying to setup the newly allocated bpath. Give
CVS a try.
What do you mean by "give cvs a try" ? I used an up to date cvs working copy,
and i can't understand how we can dereference a NULL pointer (vpath in this
case). Can you clarify your thoughts please ?
All the best,
my comment was premature. The patch I was going to propose was a bad idea.
I'm trying to track what the original code here was trying to do. It makes no
sense to populate the vpath that we are about to populate from a bpath that is
Created attachment 52402 [details] [review]
Init the bpath.
I tried the above patch. It didn't crash, but it didn't print anything though. I
think that it may be a problem with my GNOME setup. I'll do more tests as soon
I've committed a more comprehensive patch of that routine. There are other
issues in libgnomeprint that are breaking things though.