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 85153 - Segfault on exit after long session
Segfault on exit after long session
Status: RESOLVED FIXED
Product: Pan
Classification: Other
Component: general
0.12.0
Other Linux
: Normal normal
: 0.13.3
Assigned To: Charles Kerr
Charles Kerr
: 94220 101855 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2002-06-13 12:33 UTC by Carl Hudkins
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Gdb session in which I caught the segfault (5.68 KB, text/plain)
2002-06-13 12:35 UTC, Carl Hudkins
Details
Finally caught it... here's the backtrace! (20.71 KB, text/plain)
2002-06-25 02:07 UTC, Carl Hudkins
Details

Description Carl Hudkins 2002-06-13 12:33:26 UTC
Ok, this is hard to reproduce, but what usually works is to run Pan for 
several hours, downloading a lot of large multi-part binaries (such as 
videos, which typically have 9.5MB chunks, >200,000 lines apiece).  Then, 
when I close the program, I get a segfault instead of normal exit.  It 
seems like, the more I have done in the session (different groups, 
different servers, etc.) the more likely it is that Pan will crash upon 
exit, but I haven't yet found anything that will cause this 100% of the 
time.
I'll place the debug session in an attachment.
Comment 1 Carl Hudkins 2002-06-13 12:35:41 UTC
Created attachment 9194 [details]
Gdb session in which I caught the segfault
Comment 2 Charles Kerr 2002-06-13 13:59:31 UTC
There's not much to go on in this backtrace ...

1. do you have tasks running when you exit?
2. if it happens again, could you submit another backtrace?
Comment 3 Carl Hudkins 2002-06-25 02:02:54 UTC
I finally managed to make this happen again. I found that it never 
happened when I ran it under GDB, so I left out the "handle SIG..." 
stuff, because that was the only difference I could think of from 
when I run it normally.  This was a short session, in which I 
grabbed a few dozen pix and added a couple of spam/bozos.

Backtrace is attached.
Comment 4 Carl Hudkins 2002-06-25 02:07:44 UTC
Created attachment 9442 [details]
Finally caught it... here's the backtrace!
Comment 5 Charles Kerr 2002-07-15 20:24:55 UTC
Good backtrace.  The important part:

  • #0 strcmp
    from /lib/libc.so.6
  • #1 g_str_equal
    from /usr/lib/libglib-2.0.so.0
  • #2 g_hash_table_lookup
    from /usr/lib/libglib-2.0.so.0
  • #3 articlelist_get_node_from_message_id
    at articlelist.c line 1982
  • #4 articlelist_acache_changed_cb
    at articlelist.c line 2247
  • #5 pan_callback_call
    at pan-callback.c line 69
  • #6 acache_remove_entry
    at acache.c line 490
  • #7 acache_expire_to_size
    at acache.c line 615
  • #8 acache_expire_all
    at acache.c line 653
  • #9 acache_shutdown
    at acache.c line 252
  • #10 pan_shutdown
    at pan.c line 289

Though I'm not sure how this can occur.  the hash_mutex
should ensure that the strings in messageid_to_hash are
valid... hmmm...
Comment 6 Charles Kerr 2002-07-31 15:50:02 UTC
0.12 branch is probably complete, so bumping these
for investigation in 0.13.1
Comment 7 Christophe Lambin 2002-09-26 21:33:36 UTC
*** Bug 94220 has been marked as a duplicate of this bug. ***
Comment 8 Christophe Lambin 2002-10-10 23:59:49 UTC
Bumped remaining bugs to 0.13.2.
Comment 9 Christophe Lambin 2002-11-13 20:52:58 UTC
Bumped remaining bugs to 0.13.3.
Comment 10 Charles Kerr 2002-11-19 01:04:14 UTC
Fixed in cvs for 0.13.3.  I cooked up last week's speed tweaks
in articlelist to try to work around the thread issues raised in
acache & article callbacks to the header pane...

PS. "rassilon@panopticon"?
Reminds me of 0.11.2.90's release name...


Comment 11 Christophe Lambin 2003-03-30 22:46:39 UTC
*** Bug 101855 has been marked as a duplicate of this bug. ***