Created attachment 327551 [details] gecko-bugreport.txt generated with crash described Description of problem: Thunderbird crashes with attached gecko bugreport while using lightning extension. It is reproducable with certain existing events but also happened when exporting the calendar to an ICS file. Version-Release number of selected component (if applicable): [settle ~/] 4% rpm -q thunderbird thunderbird-lightning thunderbird-2.0.0.18-1.fc10.i386 thunderbird-lightning-0.9-3.fc10.i386 How reproducible: Open a specific event in my calendar to edit. not all events, just some. Also happened this morning when exporting one calendar to an ICS file. Steps to Reproduce: 1. open thunderbird and switch to calendar mode 2. select file->export calendar, select calendar, press Ok. 3. choose file name and location, select iCalendar format, press "save" 4. crash Actual results: crash with attached gecko-bugreport.txt file. Expected results: save the file in iCalendar format and return to program. Additional info: exporting to an HTML file doesn't cause the same crash.
Created attachment 327552 [details] gecko-bugreport.txt generated with crash after editing an event This gecko-bugreport.txt file was generated after editing an event and closing the window with the window manager close button. not all events cause the crash but there isn't anything special about the event that does cause the crash.
It's probably either ICAL with unsupported features or broken; but libical definitely shouldn't abort here anyways. I'm building a replacement package with aborts on such conditions disabled. Until it builds -- could you please install debugging symbols ("yum install yum-utils; debuginfo-install libical sunbird") and reproduce the problem again? We'd get a backtrace that would let us know more to find out what's wrong with your calendar. Thanks!
This is the package where the fatal aborts should be disabled, so you will probably see an error message in thunderbird, but no crash: http://koji.fedoraproject.org/koji/taskinfo?taskID=1023481
(In reply to comment #2) > It's probably either ICAL with unsupported features or broken; but libical > definitely shouldn't abort here anyways. I'm building a replacement package > with aborts on such conditions disabled. > > Until it builds -- could you please install debugging symbols ("yum install > yum-utils; debuginfo-install libical sunbird") and reproduce the problem again? > We'd get a backtrace that would let us know more to find out what's wrong with > your calendar. > > Thanks! since i haven't installed any debug packages yet this will be a ~176MB download to get installed. if this will still be useful i can do it when i have better connectivity tomorrow. please let me know.
Created attachment 327937 [details] gecko-bugreport.txt from crash after installing debuginfo rpms
(In reply to comment #4) > (In reply to comment #2) > > It's probably either ICAL with unsupported features or broken; but libical > > definitely shouldn't abort here anyways. I'm building a replacement package > > with aborts on such conditions disabled. > > > > Until it builds -- could you please install debugging symbols ("yum install > > yum-utils; debuginfo-install libical sunbird") and reproduce the problem again? > > We'd get a backtrace that would let us know more to find out what's wrong with > > your calendar. > > > > Thanks! > > since i haven't installed any debug packages yet this will be a ~176MB download > to get installed. if this will still be useful i can do it when i have better > connectivity tomorrow. please let me know. i just uploaded the gecko-bugreport.txt after the crash with debuginfo rpms (and dependencies) installed. also, i note that the libical and libical-debuginfo mentioned at: http://koji.fedoraproject.org/koji/taskinfo?taskID=1023481 are earlier versions than those from current fedora 10 repos: in koki: libical-0.32-1.fc10.nofatalerr1.i386.rpm libical-debuginfo-0.32-1.fc10.nofatalerr1.i386.rpm from fedora 10 repos: libical-0.41-1.fc10.i386 libical-debuginfo-0.41-1.fc10.i386 FYI
(In reply to comment #6) > i just uploaded the gecko-bugreport.txt after the crash with debuginfo rpms > (and dependencies) installed. Much thanks for that. I still see no backtrace there. Please check if you have gdb installed. If not, please install it and reproduce the problem again. Sorry for the inconvenience. > also, i note that the libical and libical-debuginfo mentioned at: > > http://koji.fedoraproject.org/koji/taskinfo?taskID=1023481 > > are earlier versions than those from current fedora 10 repos: > > in koki: > > libical-0.32-1.fc10.nofatalerr1.i386.rpm > libical-debuginfo-0.32-1.fc10.nofatalerr1.i386.rpm > > from fedora 10 repos: > > libical-0.41-1.fc10.i386 > libical-debuginfo-0.41-1.fc10.i386 Right; I did not notice that. I'll promptly rebuild newer packages.
http://koji.fedoraproject.org/koji/taskinfo?taskID=1025396
(In reply to comment #7) > (In reply to comment #6) > > > i just uploaded the gecko-bugreport.txt after the crash with debuginfo rpms > > (and dependencies) installed. > > Much thanks for that. I still see no backtrace there. Please check if you have > gdb installed. If not, please install it and reproduce the problem again. Sorry > for the inconvenience. i do have gdb installed: [host ~/] 65% which gdb /usr/bin/gdb i just ran it again from the command-line and the following came up when it crashed this time: ( i should have thought to check .xsession-errors - it was there as well from the previous crash. ) ---- icalerror.c:104: BADARG: Bad argument to function thunderbird-bin: icalerror.c:106: icalerror_set_errno: Assertion `0' failed. 369ec268-b38a-ec1f-1ffdc167-31d1c408 is dumped ---- also, just in case, i'll upload the gecko-bugreport.txt file from this run. > > also, i note that the libical and libical-debuginfo mentioned at: > > > > http://koji.fedoraproject.org/koji/taskinfo?taskID=1023481 > > > > are earlier versions than those from current fedora 10 repos: > > > > in koki: > > > > libical-0.32-1.fc10.nofatalerr1.i386.rpm > > libical-debuginfo-0.32-1.fc10.nofatalerr1.i386.rpm > > > > from fedora 10 repos: > > > > libical-0.41-1.fc10.i386 > > libical-debuginfo-0.41-1.fc10.i386 > > Right; I did not notice that. I'll promptly rebuild newer packages. i see the newer packages so will install those and run again.
Created attachment 327939 [details] gecko-bugreport.txt from crash after installing debuginfo rpms, second try
(In reply to comment #10) > Created an attachment (id=327939) [details] > gecko-bugreport.txt from crash after installing debuginfo rpms, second try Which packages are this? The custom built ones? Anyways, unfortunately the bug reports don't contain a backtrace. Let's see if we can get it somehow: Could you please run thunderbird as follows: $ thunderbird -g -d gdb Issue a "c" command at the (gdb) prompt and then issue a "thread apply all bt" command once thunderbird crashes? Please attach the output here.
(In reply to comment #11) > (In reply to comment #10) > > Created an attachment (id=327939) [details] [details] > > gecko-bugreport.txt from crash after installing debuginfo rpms, second try > > Which packages are this? The custom built ones? that report came from the original f10 packages - not the new custom built ones. > Anyways, unfortunately the bug reports don't contain a backtrace. Let's see if > we can get it somehow: > > Could you please run thunderbird as follows: > $ thunderbird -g -d gdb > > Issue a "c" command at the (gdb) prompt and then issue a "thread apply all bt" > command once thunderbird crashes? Please attach the output here. i installed the custom built packages and, as i'm sure you'd hoped, it doesn't crash any longer. how would you like me to proceed? i can roll back to the previous f10 shipped packages and rerun with the above commands if you'd like. just let me know what would help best.
Created attachment 338937 [details] gecko-bugreport while using internet This gecko-bugreport.txt file generats often while watching YouTube for example and closing the window with the window manager close button. not all internet sites cause the crash but it happens sometimes.
This was fixed in more recent libical packages.
This is still affecting Fedora EPEL :-(