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 309875 - galeon 1.3.21 crashed on reload on Fedora Core 4
galeon 1.3.21 crashed on reload on Fedora Core 4
Status: RESOLVED NOTGNOME
Product: galeon
Classification: Deprecated
Component: Mozilla interaction
1.3.21
Other All
: High critical
: ---
Assigned To: Philip Langdale
Yanko Kaneti
Depends on:
Blocks:
 
 
Reported: 2005-07-09 08:39 UTC by Dams
Modified: 2005-07-12 02:15 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
thread apply all bt (6.62 KB, text/plain)
2005-07-09 17:37 UTC, Dams
Details

Description Dams 2005-07-09 08:39:42 UTC
Steps to reproduce:
1. install galeon from tarball or from fedora extras rpm.
2. load a page.
3. reload it using the reload button. (insist if it doesnt crash immediatly:p)


Stack trace:
Program received signal SIGSEGV, Segmentation fault.

Thread 46912497527392 (LWP 12057)

  • #0 NSGetModule
    from /usr/lib64/mozilla-1.7.8/components/libimglib2.so
  • #1 NSGetModule
    from /usr/lib64/mozilla-1.7.8/components/libimglib2.so
  • #2 NSGetModule
    from /usr/lib64/mozilla-1.7.8/components/libnecko.so
  • #3 NSGetModule
    from /usr/lib64/mozilla-1.7.8/components/libnecko.so
  • #4 PL_HandleEvent
    from /usr/lib64/mozilla-1.7.8/libxpcom.so
  • #5 PL_ProcessPendingEvents
    from /usr/lib64/mozilla-1.7.8/libxpcom.so
  • #6 nsEventQueueImpl::~nsEventQueueImpl\$base
    from /usr/lib64/mozilla-1.7.8/libxpcom.so
  • #7 ??
    from /usr/lib64/mozilla-1.7.8/components/libwidget_gtk2.so
  • #8 g_main_context_dispatch
    from /usr/lib64/libglib-2.0.so.0
  • #9 g_main_context_check
    from /usr/lib64/libglib-2.0.so.0
  • #10 g_main_loop_run
    from /usr/lib64/libglib-2.0.so.0
  • #11 bonobo_main
    from /usr/lib64/libbonobo-2.so.0
  • #12 main
    at galeon-main.c line 281


Other information:
Bug seems to be reproducible on i386 and x86_64.
Comment 1 Dams 2005-07-09 10:48:18 UTC
Here's a more descriptive stack-trace (with gtk2/glib2/galeon/mozilla debuginfo
packages installed)

Program received signal SIGSEGV, Segmentation fault.

Thread 46912497527392 (LWP 15707)

  • #0 imgRequest::NotifyProxyListener
    at ../../../dist/include/xpcom/nsCOMPtr.h line 553
  • #1 imgCacheValidator::OnStartRequest
    at imgLoader.cpp line 917
  • #2 nsHttpChannel::HandleAsyncNotModified
    at nsHttpChannel.cpp line 416
  • #3 nsHttpChannel::AsyncCall_EventHandlerFunc
    at nsHttpChannel.cpp line 260
  • #4 PL_HandleEvent
    at plevent.c line 679
  • #5 PL_ProcessPendingEvents
    at plevent.c line 614
  • #6 nsEventQueueImpl::ProcessPendingEvents
    at nsEventQueue.cpp line 398
  • #7 event_processor_callback
    at nsAppShell.cpp line 67
  • #8 IA__g_main_context_dispatch
    at gmain.c line 1933
  • #9 g_main_context_iterate
    at gmain.c line 2564
  • #10 IA__g_main_loop_run
    at gmain.c line 2768
  • #11 bonobo_main
    at bonobo-main.c line 297
  • #12 main
    at galeon-main.c line 281

Comment 2 Philip Langdale 2005-07-09 15:34:56 UTC
It looks like a crash inside mozilla to me. There's no real galeon code in either backtrace. Note that both backtraces are not showing the same stack. You may want to try again and do a 'thread apply all bt' in gdb to get the backtrace from every thread.

But I expect that we'll close it as a mozilla bug...
Comment 3 Dams 2005-07-09 17:37:57 UTC
Created attachment 48876 [details]
thread apply all bt

Here's the backtrace of all threads. There doesnt seem to be anything
interesting in the "other threads"... This is not the same stack trace again,
but this is always triggered by the same action (click on reload button...),
how can that be ? Could that depend of the number of tabs opened ? is that some
kind of buffer overflow ? or more likely a race condition, no ?
Comment 4 Philip Langdale 2005-07-09 19:05:37 UTC
Well, it's pretty much the same backtrace as in comment #1. I'm pretty sure that galeon doesn't attach any listeners that could end up getting notified by the imgRequest, so I really fail to see how our code is doing anything.

You could try running TestGtkEmbed or epiphany and see what that does but I suspect they don't bother including TestGtkEmbed in the mozilla/firefox packages in fc4.

If you get a crash with one of those, then it's pretty clearly a mozilla problem, even if it doesn't happen with the mozilla-browser/firefox
Comment 5 Robin Humble 2005-07-11 19:22:06 UTC
I get an immediate crash when reloading on epiphany on x86_64... err, and
actually with mozilla itself too! so that makes it a mozilla bug.

according to this thread:
  https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=160330
it's probably this mozilla bug:
  https://bugzilla.mozilla.org/show_bug.cgi?id=293307
Comment 6 Philip Langdale 2005-07-12 02:15:42 UTC
Then our work here is done. Thanks for following up.