Description of problem: With the new Firefox 1.0.7 update for FC4, I'm getting Firefox crashing randomly. Running inside gdb, I'm getting the following backtrace all the time: (gdb) bt #0 0x00f69402 in __kernel_vsyscall () #1 0x00921a1c in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib/libpthread.so.0 #2 0x001525a1 in ?? () #3 0x0848d614 in ?? () #4 0x0848d5b0 in ?? () #5 0xb34463dc in ?? () #6 0x0000ea60 in JS_HashTableEnumerateEntries@plt () #7 0x00152880 in ?? () #8 0x0000003c in ?? () #9 0x07862d18 in ?? () from /usr/lib/firefox-1.0.7/components/libnecko.so #10 0x0000ea60 in JS_HashTableEnumerateEntries@plt () #11 0x0775b9db in NSGetModule () from /usr/lib/firefox-1.0.7/components/libnecko.so #12 0x00156916 in ?? () #13 0x0848d580 in ?? () #14 0x00928ff4 in ?? () from /lib/libpthread.so.0 #15 0x00000000 in ?? () Version-Release number of selected component (if applicable): firefox-1.0.7-1.1.fc4 How reproducible: Right now I can reproduce it whenever I try to submit an attachment to gnome bugzilla, but it has been crashing at other times too. Additional info: With the same update, I and a friend of mine have been observing problems with font selection too, but that probably is another problem.
Another backtrace, crashing consistently when pressing the browse button in bugzill: Program received signal SIG33, Real-time event 33. [Switching to Thread -1222104144 (LWP 19958)] 0x00d42402 in __kernel_vsyscall () (gdb) (gdb) bt #0 0x00d42402 in __kernel_vsyscall () #1 0x00921a1c in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib/libpthread.so.0 #2 0x006d25a1 in PR_Lock () from /usr/lib/firefox-1.0.7/libnspr4.so #3 0x006d2880 in PR_WaitCondVar () from /usr/lib/firefox-1.0.7/libnspr4.so #4 0x00d569db in NSGetModule () from /usr/lib/firefox-1.0.7/components/libnecko.so #5 0x006d6916 in PR_Select () from /usr/lib/firefox-1.0.7/libnspr4.so #6 0x0091fb80 in start_thread () from /lib/libpthread.so.0 #7 0x005bd9ce in clone () from /lib/libc.so.6
I'm using the same version of Firefox and it also crashes when pushing a "browse" button, but it only happens a very few times. I will try the get some output.
I have been experiencing the same random "disappearing" with Firefox and OpenOffice.org. Over at Bug #171873 there is a report that a fix was found. Is it at all possible that a similar bug found its way into Ooo that is still in Firefox?
For what it's worth, I'm not experimenting that anymore. Guess something was very wrong with my Gtk+.
Can you be more specific about what you thought was wrong with your Gtk+, and suggestions that other people can try who still see this problem? Note: With OOo they seem to think it may be a compiler optimization bug...
Don't exactly remember, but after a couple of days it never happened again.