Bug 659411 - [abrt] g_type_check_instance_cast in progress bar drawing (SIGSEGV)
Summary: [abrt] g_type_check_instance_cast in progress bar drawing (SIGSEGV)
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Fedora
Classification: Fedora
Component: libreoffice
Version: rawhide
Hardware: x86_64
OS: Unspecified
low
medium
Target Milestone: ---
Assignee: Caolan McNamara
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: abrt_hash:a3377c9a70982247de5ae1a8dd8...
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-12-02 18:29 UTC by Horst H. von Brand
Modified: 2011-02-09 16:45 UTC (History)
3 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2011-02-09 16:45:24 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
File: backtrace (49.59 KB, text/plain)
2010-12-02 18:29 UTC, Horst H. von Brand
no flags Details
Yum Log for Nov/Dec 2010 (138.07 KB, application/octet-stream)
2010-12-05 14:45 UTC, Stephen
no flags Details
yum.log for dec 01 through dec 03 (13.69 KB, application/octet-stream)
2010-12-05 23:50 UTC, Horst H. von Brand
no flags Details

Description Horst H. von Brand 2010-12-02 18:29:10 UTC
abrt version: 1.1.14
architecture: x86_64
Attached file: backtrace
cmdline: /usr/lib64/libreoffice/program/soffice.bin -quickstart -nologo -nodefault
comment: I had edited this file successfully before, printed it out (it is a form) and then exited oocalc without saving. Now I tried to open it again.
component: libreoffice
executable: /usr/lib64/libreoffice/program/soffice.bin
kernel: 2.6.36.1-10.fc15.x86_64
package: libreoffice-core-3.2.99.3-2.fc15
reason: Process /usr/lib64/libreoffice/program/soffice.bin was killed by signal 11 (SIGSEGV)
release: Fedora release 15 (Rawhide)
time: 1291314393
uid: 500

How to reproduce
-----
1. oocalc /tmp/Rec\ Gastos.xls
2.
3.

Comment 1 Horst H. von Brand 2010-12-02 18:29:12 UTC
Created attachment 464323 [details]
File: backtrace

Comment 2 Horst H. von Brand 2010-12-02 18:31:03 UTC
Tried opening the exact same file again (after trying in gnumeric), now it worked.

Comment 3 Caolan McNamara 2010-12-02 20:07:58 UTC
Hmm, there's been a lot of those g_type_check_instance_cast crashes in the past. I think this is the first time we've gotten enough of a stacktrace to indicate that its the progressbar that's involved. 

Doesn't really help me to get closer to what's causing this.

Comment 4 Caolan McNamara 2010-12-02 20:35:00 UTC
Could you attach your /var/log/yum.log. I wonder if this happens if gtk2 or glib or something gets upgraded while the (cursed) quickstarter is running.

Comment 5 Stephen 2010-12-05 13:17:38 UTC
Package: libreoffice-core-3.2.99.3-2.fc15
Architecture: x86_64
OS Release: Fedora release 15 (Rawhide)


How to reproduce
-----
1. Attempted to start libreoffice using the command libreoffice -calc.  Nothing happened
2. Tried the command libreoffice on its own
3. Crash

Comment 6 Caolan McNamara 2010-12-05 13:26:09 UTC
caolanm->stephen: Same question to you. Did you just do a yum update or otherwise upgrade your packages while logged into the desktop ?

i.e. did this happen because the quickstarter was running and something got upgraded while it was running. Please attach your /var/log/yum.log if you can.

Comment 7 Stephen 2010-12-05 14:45:10 UTC
Created attachment 464852 [details]
Yum Log for Nov/Dec 2010

Attached is the YUM log entries for Nov and December this year.  I almost certainly did a Yum update whilst LibreOffice was running

Comment 8 Caolan McNamara 2010-12-05 16:20:44 UTC
I wonder if upgrading "gtk-theme-engine-clearlooks" while running triggers this. Its a possibility at least

Comment 9 Horst H. von Brand 2010-12-05 23:50:38 UTC
Created attachment 464905 [details]
yum.log for dec 01 through dec 03

There were no updates installed on dec 02 (the date of the crash), and I definitely had rebooted the machine after the update.

Comment 10 Horst H. von Brand 2010-12-06 20:31:40 UTC
Package: libreoffice-core-3.2.99.3-2.fc15
Architecture: x86_64
OS Release: Fedora release 15 (Rawhide)


How to reproduce
-----
1. Closed oocalc's window
2.
3.


Comment
-----
Had a window open for a while (was away from the machine), and tried to close via X on it; a dialog "not responding" opened, declined to kill immedietaly

Comment 11 Caolan McNamara 2011-02-09 16:45:24 UTC
This one baffles me, crash is in drawing the progress bar, but the theme is a standard one so its not simply that, might be due to an update while running, but #9 rules that out so leaves me with a heisenbug where the real cause is likely something totally unrelated to the final point of crash :-(

If you still have the .xls which *might* have triggered this feel free to attach it here and I'll have a last-ditch import of it under valgrind to see if its anyway related.


Note You need to log in before you can comment on or make changes to this bug.