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 361804 - Metacity crashed on changing workspaces from 16 to 36
Metacity crashed on changing workspaces from 16 to 36
Status: RESOLVED FIXED
Product: metacity
Classification: Other
Component: general
2.20.x
Other Linux
: Normal critical
: ---
Assigned To: Metacity maintainers list
Metacity maintainers list
Depends on:
Blocks:
 
 
Reported: 2006-10-12 21:19 UTC by Sebastien Bacher
Modified: 2008-07-13 22:22 UTC
See Also:
GNOME target: ---
GNOME version: 2.19/2.20

Attachments
I think this patch (against trunk, shouldn't be hard to backport) will fix the crash (1.56 KB, patch)
2007-04-09 20:55 UTC, Elijah Newren 		committed Details | Review

Description Sebastien Bacher 2006-10-12 21:19:19 UTC 
That bug has been opened on https://launchpad.net/distros/ubuntu/+source/metacity/+bug/64818

"On increasing the number of workspaces from 16 to 36, metacity crashed.

steps :

1. In gnome panel, use Workspace Switcher Applet
2. Go to preferences, and increase num from 16 to 36
3. metacity crashes

http://librarian.launchpad.net/4707965/_usr_bin_metacity.99999.crash
Crash file
...
ExecutablePath: /usr/bin/metacity
Package: metacity 1:2.16.3-0ubuntu2
ProcCmdline: /usr/bin/metacity --sm-client-id=default0
...
> Thanks for your bug. How did you change the number of workspace, with the keyboard or by clicking? Could you install metacity-dbgsym from "deb http://people.ubuntu.com/~pitti/ddebs edgy main" and get a backtrace with it?
...
I changed it by clicking on the up-arrow.

BTW, I saw the crash again while scaling down from 36 to 16 workspaces.
...
http://librarian.launchpad.net/4794019/Stacktrace
apport-retrace backtrace
...Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1".
Core was generated by `/usr/bin/metacity --sm-client-id=default0'.
Program terminated with signal 11, Segmentation fault.

+ Trace 76050

  • #0 meta_workspace_free
    at workspace.c line 121
  • #0 meta_workspace_free
    at workspace.c line 121
  • #1 update_num_workspaces
    at screen.c line 1156
  • #2 changed_idle_handler
    at prefs.c line 257
  • #3 g_source_is_destroyed
    from /usr/lib/libglib-2.0.so.0
  • #4 g_main_context_dispatch
    from /usr/lib/libglib-2.0.so.0
  • #5 g_main_context_check
    from /usr/lib/libglib-2.0.so.0
  • #6 g_main_loop_run
    from /usr/lib/libglib-2.0.so.0
  • #7 main
    at main.c line 389
  • #8 __libc_start_main
    from /lib/tls/i686/cmov/libc.so.6
  • #9 _start 






Comment 1


Sebastien Bacher





          2006-12-06 08:49:32 UTC
        



The submitter doesn't get the problem with edgy (GNOME 2.16.1), I've closed the Ubuntu bug, you might want to close that bug too






Comment 2


Sebastien Bacher





          2006-12-21 13:32:03 UTC
        



New Ubuntu bug with edgy: https://launchpad.net/distros/ubuntu/+source/metacity/+bug/76203






Comment 3


Elijah Newren





          2006-12-21 17:24:11 UTC
        



Sebastien: How is ubuntu #76203 a duplicate of ubuntu #64818?  The former was about crash on launch while the latter was about crash when changing the number of workspaces.  I didn't notice any other information that would suggest the two are duplicates, though perhaps I missed something?






Comment 4


Sebastien Bacher





          2006-12-21 23:08:10 UTC
        



The backtrace from #76203 looks the same (using apport-retrace with debug packages installed to get a debug backtrace):

"...




+
Trace
    96094






    

  • 
#0
meta_workspace_free

    
                at workspace.c
                  line
                  121

    

    • 

  • 
#1
update_num_workspaces

    
                at screen.c
                  line
                  1156

    

    • 

  • 
#2
changed_idle_handler

    
                at prefs.c
                  line
                  257

    

    • 

  • 
#3
g_idle_dispatch

    
                at gmain.c
                  line
                  3926

    

    • 

  • 
#4
IA__g_main_context_dispatch

    
                at gmain.c
                  line
                  2045

    

    • 

  • 
#5
g_main_context_iterate

    
                at gmain.c
                  line
                  2677

    

    • 

  • 
#6
IA__g_main_loop_run

    
                at gmain.c
                  line
                  2881

    

    • 

  • 
#7
main

    
                at main.c
                  line
                  389

    

    • 

  • 
#8
__libc_start_main

    
                from
                /lib/tls/i686/cmov/libc.so.6

    

    • 

  • 
#9
_start

    • 








The crashes are stored to /var/crash and an icon is displayed to the notification area when a crash file has been placed there. The description looks like that the user noticed the icon signaling the crash after login, not that the crash itself happened then






Comment 5


Sebastien Bacher





          2007-02-20 09:36:19 UTC
        



New ubuntu bug about that with GNOME 2.17: https://launchpad.net/ubuntu/+source/metacity/+bug/86436






Comment 6


Elijah Newren





          2007-04-09 20:55:59 UTC
        



Created attachment 86071 [details] [review]
I think this patch (against trunk, shouldn't be hard to backport) will fix the crash

Trying to look for something that could cause a crash with this stack trace, I think the above patch catches the problem (trying to free an array of lists as well as the lists, when they've already been freed).  Any chance you could test and verify?  For some reason, I had a hard time duplicating.






Comment 7


Sebastien Bacher





          2007-09-27 15:51:31 UTC
        



Duplicate on GNOME 2.20 https://bugs.launchpad.net/ubuntu/+source/metacity/+bug/144527






Comment 8


Thomas Thurman





          2008-07-13 22:22:58 UTC
        



The patch is common sense, and causes no apparent problems, so I've committed it.  Let's see whether we get anyone reporting it again.  Until then, unless anyone can tell us how to reproduce the problem so we can say for sure, I think we can tentatively say it's fixed.