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 364839 - Many applications crash on exit() when GNOME Accessibility is in use (especially in FC5 and FC6 currently)
Many applications crash on exit() when GNOME Accessibility is in use (especia...
Status: RESOLVED FIXED
Product: at-spi
Classification: Platform
Component: atkbridge
unspecified
Other All
: Normal critical
: ---
Assigned To: Li Yuan
Li Yuan
: 356266 388053 388062 388450 389327 389723 390111 390362 390403 390533 390632 390901 390929 390946 391143 391679 391697 391772 391862 391885 392005 392121 392138 393083 393096 393322 393537 393558 393587 393593 393725 394024 394056 394123 394443 394757 394937 395006 395218 395220 395679 395753 396421 405261 450330 476390 (view as bug list)
Depends on:
Blocks: 391331
 
 
Reported: 2006-10-24 22:02 UTC by Jouni.Lohikoski
Modified: 2007-09-14 15:22 UTC
See Also:
GNOME target: ---
GNOME version: 2.13/2.14


Attachments
patch should fix this bug (1.86 KB, patch)
2006-12-28 03:37 UTC, Li Yuan
committed Details | Review

Description Jouni.Lohikoski 2006-10-24 22:02:35 UTC
Steps to reproduce:
When trying to change any preferences in System/Preferences/* something
crashes when the change has been made and the dialog window is closed.
System log gets this:

Oct 24 23:28:01 localhost kernel: gnome_segv2[3072]: segfault at
0000000000000000 rip 000000336d873cb0 rsp 00007fff8634f688 error 4

Steps to reproduce the crash:
1.  Allow other hosts connect gdm with XDMCP
(/etc/gdm/custom.conf  <-  "[xdmcp] Enable=true") (open a hole to the
firewall for port 177)

2.  Connect in from another host:
$ X -query <gdm-host-IP-address>

3. Log in

4. Try to change for example mouse preferences: System | Preferences |
Mouse. (Or, System | About_GNOME)

5. Click OK in the dialog window => crash

Expected Results:
The wanted change in the preferences would be made and nothing would
crash.

How often does this happen?
Every time.

Additional Information:
System log of the gdm-host gets always something like this:
Oct 24 23:28:01 localhost kernel: gnome_segv2[3072]: segfault at
0000000000000000 rip 000000336d873cb0 rsp 00007fff8634f688 error 4


Stack trace:
Backtrace was generated from '/usr/bin/gnome-about'

(no debugging symbols found)
Using host libthread_db library "/lib64/libthread_db.so.1".
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
....
[Thread debugging using libthread_db enabled]
[New Thread 46912496320080 (LWP 3606)]
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
...
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
0x000000336f30ccb5 in waitpid () from /lib64/libpthread.so.0

Thread 1 (Thread 46912496320080 (LWP 3606))

  • #0 waitpid
    from /lib64/libpthread.so.0
  • #1 gnome_gtk_module_info_get
    from /usr/lib64/libgnomeui-2.so.0
  • #2 <signal handler called>
  • #3 strlen
    from /lib64/libc.so.6
  • #4 giop_send_buffer_append_string
    from /usr/lib64/libORBit-2.so.0
  • #5 ORBit_marshal_value
    from /usr/lib64/libORBit-2.so.0
  • #6 ORBit_marshal_value
    from /usr/lib64/libORBit-2.so.0
  • #7 ORBit_marshal_any
    from /usr/lib64/libORBit-2.so.0
  • #8 ORBit_marshal_value
    from /usr/lib64/libORBit-2.so.0
  • #9 ORBit_marshal_value
    from /usr/lib64/libORBit-2.so.0
  • #10 ORBit_small_allocbuf
    from /usr/lib64/libORBit-2.so.0
  • #11 ORBit_small_invoke_stub
    from /usr/lib64/libORBit-2.so.0
  • #12 Accessibility_EventListener_notifyEvent
    from /usr/lib64/libspi.so.0
  • #13 gnome_accessibility_module_shutdown
    from /usr/lib64/gtk-2.0/modules/libatk-bridge.so
  • #14 gnome_accessibility_module_shutdown
    from /usr/lib64/gtk-2.0/modules/libatk-bridge.so
  • #15 exit
    from /lib64/libc.so.6
  • #16 __libc_start_main
    from /lib64/libc.so.6
  • #17 ??
  • #18 ??
  • #19 ??


Other information:
All System | Preferences dialog windows will crash also.

This doesn't happen, when the same system (gdm/GNOME/metacity) is used from local console.
Comment 1 Vincent Untz 2006-11-13 17:46:55 UTC
Could this be an a11y issue? (see stack trace)
Comment 2 bill.haneman 2006-11-13 18:04:09 UTC
Definitely not ATK (see stack trace); possibly at-spi.
Possibly exposed by the new "remote apps accessibility" patch to at-spi? Is the reporter certain that the gnome env in question is 2.13/2.14?
Comment 3 bill.haneman 2006-11-14 16:50:47 UTC
Aha, this happens after exit() has been called.  Lowering severity (crash on exit).

Could be an ORBit2 thing?  
Comment 4 Jouni.Lohikoski 2006-11-16 19:22:41 UTC
> Is the reporter certain that the gnome env in question is 2.13/2.14?

Yes. The desktop´s Panel-menu: System/About_GNOME says "2.14.3", and when "Close"-button is pushed, it crashes.
Comment 5 Jouni.Lohikoski 2006-11-16 19:28:36 UTC
Forgot to mention, the OS is to the formatted/empy disk installed Fedora core 5, with updates until this date.
The X-server (the remote X-terminal) is fc3-installation w/ updates.
Comment 6 Derek Atkins 2006-12-27 16:00:31 UTC
For the record this happens on my FC5 system even with a local display.. And it started after I turned on Accessibility.  GnuCash crashes (locally) with a very similar stack trace (after exit).  If I turn off accessibility, the crash goes away.  So it has nothing to do with remote display, but it does seem to be an issue in FC5's gnome/at-spi libraries.
Comment 7 Pedro de Medeiros 2006-12-27 18:54:19 UTC
By the look of it, it is yet another duplicate of bug 389327, but this report has more information, so setting bug 389327 as a duplicate of this.
Comment 8 Pedro de Medeiros 2006-12-27 18:54:43 UTC
*** Bug 389327 has been marked as a duplicate of this bug. ***
Comment 9 Pedro de Medeiros 2006-12-27 18:57:14 UTC
*** Bug 388053 has been marked as a duplicate of this bug. ***
Comment 10 Pedro de Medeiros 2006-12-27 18:58:54 UTC
*** Bug 388062 has been marked as a duplicate of this bug. ***
Comment 11 Pedro de Medeiros 2006-12-27 19:00:58 UTC
*** Bug 388450 has been marked as a duplicate of this bug. ***
Comment 12 Pedro de Medeiros 2006-12-27 19:02:08 UTC
*** Bug 389723 has been marked as a duplicate of this bug. ***
Comment 13 Li Yuan 2006-12-28 03:18:22 UTC
I think the real trace is http://bugzilla.gnome.org/attachment.cgi?id=78964&action=view
Comment 14 Li Yuan 2006-12-28 03:37:18 UTC
Created attachment 78972 [details] [review]
patch should fix this bug
Comment 15 Jens Granseuer 2006-12-29 16:50:36 UTC
*** Bug 390533 has been marked as a duplicate of this bug. ***
Comment 16 Jens Granseuer 2006-12-29 17:35:23 UTC
*** Bug 390632 has been marked as a duplicate of this bug. ***
Comment 17 Jens Granseuer 2006-12-29 17:38:09 UTC
*** Bug 390403 has been marked as a duplicate of this bug. ***
Comment 18 Jens Granseuer 2006-12-30 10:05:21 UTC
*** Bug 390901 has been marked as a duplicate of this bug. ***
Comment 19 Jens Granseuer 2006-12-30 15:11:06 UTC
*** Bug 390946 has been marked as a duplicate of this bug. ***
Comment 20 Jens Granseuer 2006-12-30 16:38:31 UTC
*** Bug 390929 has been marked as a duplicate of this bug. ***
Comment 21 Saikat Guha 2006-12-30 21:51:18 UTC
*** Bug 390111 has been marked as a duplicate of this bug. ***
Comment 22 Susana 2006-12-31 02:46:15 UTC
*** Bug 391143 has been marked as a duplicate of this bug. ***
Comment 23 Susana 2007-01-02 12:49:38 UTC
*** Bug 391772 has been marked as a duplicate of this bug. ***
Comment 24 palfrey 2007-01-02 18:38:27 UTC
*** Bug 391885 has been marked as a duplicate of this bug. ***
Comment 25 palfrey 2007-01-02 18:38:36 UTC
*** Bug 391862 has been marked as a duplicate of this bug. ***
Comment 26 Jonathon Jongsma 2007-01-02 20:51:52 UTC
*** Bug 392005 has been marked as a duplicate of this bug. ***
Comment 27 Jens Granseuer 2007-01-02 22:07:08 UTC
*** Bug 391679 has been marked as a duplicate of this bug. ***
Comment 28 Jens Granseuer 2007-01-02 22:08:20 UTC
*** Bug 391697 has been marked as a duplicate of this bug. ***
Comment 29 Benoît Dejean 2007-01-03 11:11:17 UTC
*** Bug 392138 has been marked as a duplicate of this bug. ***
Comment 30 Bruno Boaventura 2007-01-03 17:18:41 UTC
*** Bug 392121 has been marked as a duplicate of this bug. ***
Comment 31 Christian Kirbach 2007-01-05 11:44:57 UTC
*** Bug 393096 has been marked as a duplicate of this bug. ***
Comment 32 Bruno Boaventura 2007-01-05 17:07:03 UTC
*** Bug 393083 has been marked as a duplicate of this bug. ***
Comment 33 Christian Kirbach 2007-01-05 23:55:48 UTC
*** Bug 393322 has been marked as a duplicate of this bug. ***
Comment 34 Jens Granseuer 2007-01-06 18:06:21 UTC
*** Bug 393537 has been marked as a duplicate of this bug. ***
Comment 35 Jens Granseuer 2007-01-06 22:00:10 UTC
*** Bug 393593 has been marked as a duplicate of this bug. ***
Comment 36 Jens Granseuer 2007-01-06 22:00:47 UTC
*** Bug 393587 has been marked as a duplicate of this bug. ***
Comment 37 Jens Granseuer 2007-01-06 22:02:11 UTC
*** Bug 393558 has been marked as a duplicate of this bug. ***
Comment 38 Jens Granseuer 2007-01-06 22:03:02 UTC
*** Bug 390362 has been marked as a duplicate of this bug. ***
Comment 39 Jens Granseuer 2007-01-07 20:25:16 UTC
*** Bug 393725 has been marked as a duplicate of this bug. ***
Comment 40 Gabor Kelemen 2007-01-08 12:44:48 UTC
*** Bug 394123 has been marked as a duplicate of this bug. ***
Comment 41 Vincent Untz 2007-01-08 14:04:36 UTC
*** Bug 394056 has been marked as a duplicate of this bug. ***
Comment 42 Gabor Kelemen 2007-01-08 23:56:24 UTC
*** Bug 394443 has been marked as a duplicate of this bug. ***
Comment 43 Jim Cornette 2007-01-09 02:02:51 UTC
So I gather that ORBit2 is where the problem resides. What is the fix? Comment #3 references possibly ORBit2 as suspect.

Questioning because of Bug 394443 being marked as Resolved and a duplicate of this bug.
Comment 44 Li Yuan 2007-01-09 06:43:51 UTC
You can see my patch below. The bug will be fixed in at-spi 1.7.15 with gnome 2.17.5.
Comment 45 Jim Cornette 2007-01-09 23:21:54 UTC
So are spi_atk_... functions causing the crashes on exit that other applications are exhibiting?
Thanks for the explanation. I do not work with code so I'll wait for the releases that you referenced above. I have at-spi-1.7.14-1.fc7 installed right now.
Comment 46 Li Yuan 2007-01-10 02:30:21 UTC
Yes, this will be fixed in at-spi-1.7.15.
Comment 47 palfrey 2007-01-11 01:03:57 UTC
*** Bug 395006 has been marked as a duplicate of this bug. ***
Comment 48 palfrey 2007-01-11 01:04:08 UTC
*** Bug 395218 has been marked as a duplicate of this bug. ***
Comment 49 palfrey 2007-01-11 01:04:16 UTC
*** Bug 395220 has been marked as a duplicate of this bug. ***
Comment 50 palfrey 2007-01-11 01:04:25 UTC
*** Bug 394937 has been marked as a duplicate of this bug. ***
Comment 51 palfrey 2007-01-11 01:04:30 UTC
*** Bug 394757 has been marked as a duplicate of this bug. ***
Comment 52 Li Yuan 2007-01-13 03:13:32 UTC
*** Bug 394024 has been marked as a duplicate of this bug. ***
Comment 53 palfrey 2007-01-15 00:24:48 UTC
*** Bug 396421 has been marked as a duplicate of this bug. ***
Comment 54 Vincent Untz 2007-01-15 19:57:56 UTC
*** Bug 395679 has been marked as a duplicate of this bug. ***
Comment 55 Vincent Untz 2007-01-17 18:27:21 UTC
*** Bug 395753 has been marked as a duplicate of this bug. ***
Comment 56 Jim Cornette 2007-01-18 02:53:42 UTC
Am I understanding that Fedora Development is now fixed but the problem with applications crashing on exit are happening in FC5 and FC6?

Anyway, the patch works and upgrading to at-spi-1.7.15-1.fc7 stops the problem with applications crashing on exit.

A downstream bug report should probably be filed for the FC6 and FC5 problem if true.
Comment 57 Francesco Mari 2007-02-07 12:27:25 UTC
*** Bug 405261 has been marked as a duplicate of this bug. ***
Comment 58 Karsten Bräckelmann 2007-02-07 12:46:11 UTC
*** Bug 356266 has been marked as a duplicate of this bug. ***
Comment 59 Vincent Untz 2007-06-26 21:57:02 UTC
*** Bug 450330 has been marked as a duplicate of this bug. ***
Comment 60 Cosimo Cecchi 2007-09-14 15:22:37 UTC
*** Bug 476390 has been marked as a duplicate of this bug. ***