GNOME Bugzilla – Bug 778419
Crash in icaltimezone_expand_changes
Last modified: 2017-11-24 22:10:41 UTC
Found in coredumpctl:
+ Trace 237129
Bug #634294 has a similar trace, but that bug is closed.
What are your versions of gnome-calendar, libical and evolution-data-server, please? I do not see anything like that filled above. Also, is this reproducible, or just a "random" crash?
It's a random crash. I haven't actually run gnome-calendar is several weeks; seems it must be running and crashing itself in the background. gnome-calendar-3.22.1-1.fc25 libical-2.0.0-8.fc25 evolution-data-server-3.22.4-1.fc25
Is this reproducible against latest master and/or the 3.24 stable release?
I don't see recent reports of this crash specifically, but here are a bunch of similar crashes with tons of reports: https://bugzilla.redhat.com/show_bug.cgi?id=1399502 https://bugzilla.redhat.com/show_bug.cgi?id=1409398 https://bugzilla.redhat.com/show_bug.cgi?id=1427995 https://bugzilla.redhat.com/show_bug.cgi?id=1414464
Basically gnome-calendar is still *really* crashy compared to other GNOME apps. Lots of crashes in libical. It's compounded by the fact that calendar is installed by default and running in the background at all times. If you don't notice it on Arch, that's probably just because you don't get crash notifications like we do on Fedora. Check coredumpctl for gnome-calendar crashes and I'm sure you'll see a fair few.
(In reply to Michael Catanzaro from comment #6) > Basically gnome-calendar is still *really* crashy compared to other GNOME > apps. Lots of crashes in libical. That's really sad. I have quite a different experience on Arch. > If you don't notice it on Arch, that's probably just because you don't get crash notifications like we do on Fedora. I don't get the crashes. At all. > Check coredumpctl for gnome-calendar crashes and I'm sure you'll see a fair few. I often do that. Calendar is rock-solid on Arch. It's also sad the fact that all those coredumps and other valuable information are all stored in a downstream bugtracker, and none of them were linked here until now. I don't use Fedora, and I don't watch Red Hat's bugtracker. That's why we have a GNOME Bugzilla instance running. If any one of those RH tickets were linked here (or even better, if a ~single~ coredump like those), I'd be able to fix that much quicker.
I'd recommend keeping an eye on FAF reports. E.g.: https://retrace.fedoraproject.org/faf/problems/?component_names=gnome-calendar&associate=__None&daterange=2017-04-03:2017-04-17&bug_filter=None&function_names=&binary_names=&source_file_names=&since_version=&since_release=&to_version=&to_release=# shows crashes that were reported in the last two weeks. Some of the reports are not very useful, like that first one that's clearly a bunch of unrelated reports incorrectly lumped together. But the second crash (icaltimezone_ensure_coverage, 573 reports) looks actionable. Also those four crashes in icaltimezone_compare_change_fn (182 + 134 + 70 + 35). I can get you a full backtrace from any of the private bugs on request, but the truncated backtraces provided by FAF are often sufficient. If you scroll down you'll see more crashes in gcal_manager_source_enabled, gcal_manager_is_client_writable, icaltime_day_of_year... a search provider crash in gcal-utils get_circle_pixbuf_from_color... lots and lots of interesting reports here, but indeed there's no mechanism to report them upstream.
From my semi-recent experience on similar libical-related crashes, it can happen due to using already freed libical structures, like happened in bug #634294 (its change is part of 3.22.0). It could also be that the timezone structures were freed incorrectly, like with this commit [1] in eds (3.16.0+). Thus it's nothing known yet, but it can be in eds too. [1] https://git.gnome.org/browse/evolution-data-server/commit/?id=260b0f229b76
-- GitLab Migration Automatic Message -- This bug has been migrated to GNOME's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.gnome.org/GNOME/gnome-calendar/issues/111.