Description of problem: Firefox crashes randomly on me when using piotripawel's web site for internet shopping. Version-Release number of selected component (if applicable): firefox-1.5.0.9-2.fc6 gnash-0.7.2-1.mikedlr How reproducible: very random. Happens quite often when I am actually shopping and do product searches. During testing I could reproduce it with strace running and not at all with --verbose --sync options to FireFox. Steps to Reproduce: 0. Create a completely fresh user. 1. log in to http://www.piotripawel.pl/e-zakupy/ 2. choose "Warszawa" and press "WEJśCIE 3. type things into the "ZNAJDż TORWAR" box (search) such as "piwo", "wino", "ser" 4. select items to buy / go backwards. Actual results: mozilla crashes randomly. Expected results: mozilla is stable and fine. or mozilla gives a clear error message. Additional info: this appears to be a heisenbug. Turning on debugging output from firefox makes it go away.. maybe timing related? I have a build of gnash which I did myself; see Bug #217020 however a) it's just a rebuild of the redhat srpm. b) gnash crashing should at least have firefox give an error message. Recommendations for troubleshooting appreciated; particularly how to get a proper backtrace.
Created attachment 147858 [details] strace of firefox crashing
Created attachment 147859 [details] log of attempting and failing to get a backtrace showing X errors running "firefox -g -d gdb" I managed to see some X errors. No backtrace since the program seems to exit normally.
Cannot reprorduce with flash-plugin-9.0.31.0-1 (half-liter of the Czech Budweiser is terribly expensive, which IMHO is much important bug than that gnash is broken ;-)), which IMHO means that this is bug in gnash. Changing component accordingly.
After discussion with FF people closing this bug as WORKSFORME. If you can reproduce it using our own packages, then please reopen with additional information. As far as recommendation for troubleshooting, it probably really belongs to somewhere else than to our Bugzilla.
Sounds like a gnash issue anyway.
(In reply to comment #5) > Sounds like a gnash issue anyway. That's what I meant -- especially with custom-made gnash. We really cannot support that.
(In reply to comment #6) > (In reply to comment #5) > > Sounds like a gnash issue anyway. Not completely sure. There could also be a firefox or an X bug somewhere. Agreed, a bug in gnash is the best guess. > That's what I meant -- especially with custom-made gnash. We really cannot > support that. I can't see why. The rpm from fedora is unusable on some systems while a rebuild leads to something that works. See Bug 217020. So in this case I think it is acceptable to support a custom-made gnash done by recompiling the srpm. However I cannot reproduce that bug because I don't understand the polish language.
(In reply to comment #7) > I can't see why. The rpm from fedora is unusable on some systems while > a rebuild leads to something that works. See Bug 217020. So in this case I > think it is acceptable to support a custom-made gnash done by recompiling > the srpm. OK, whatever. If you want, it is your package, so do whatever you feel right. Reopen it, dance with it, really whatever. :-) > However I cannot reproduce that bug because I don't understand the polish > language. You don't have to -- yes, I can understand Polish (being a Czech), but I think the instruction above are pretty usable even without any real understanding of Polish.
(In reply to comment #8) > (In reply to comment #7) > > I can't see why. The rpm from fedora is unusable on some systems while > > a rebuild leads to something that works. See Bug 217020. So in this case I > > think it is acceptable to support a custom-made gnash done by recompiling > > the srpm. > > OK, whatever. If you want, it is your package, so do whatever you feel right. > Reopen it, dance with it, really whatever. :-) ;-) > > However I cannot reproduce that bug because I don't understand the polish > > language. > > You don't have to -- yes, I can understand Polish (being a Czech), but I think > the instruction above are pretty usable even without any real understanding of > Polish. I don't want to add myself as a user when I don't understand the conditions...
The gnash 7.2 version isn't supported by the gnash developers, so it would be convenient if the CVS version could be tried to see if it is reproducible. I don't think it is really worth to fix the gnash 7.2 version which is the fedora version since the upstream gnash has already changed a lot. If this is an issue present in CVS and there is a common fix, however I'll do the fix.