It seems like many different GNOME shell crashes are reported (by abrt/libreport) as dupes of https://bugzilla.redhat.com/show_bug.cgi?id=1469813 , while they shouldn't be. It seems like this happens when the backtrace starts like this: #0 _g_log_abort at gmessages.c:549 #1 g_log_default_handler at gmessages.c:3039 #2 default_log_handler at main.c:303 Here are all the bugs that were initially considered dupes by abrt: https://bugzilla.redhat.com/show_bug.cgi?id=1508398 https://bugzilla.redhat.com/show_bug.cgi?id=1507514 https://bugzilla.redhat.com/show_bug.cgi?id=1507099 https://bugzilla.redhat.com/show_bug.cgi?id=1506325 https://bugzilla.redhat.com/show_bug.cgi?id=1506257 https://bugzilla.redhat.com/show_bug.cgi?id=1502183 https://bugzilla.redhat.com/show_bug.cgi?id=1501396 https://bugzilla.redhat.com/show_bug.cgi?id=1498261 If you look at them, they all have those first three frames in their backtraces, but beyond that the backtraces differ, and other details about the bugs clearly differ also. I *think* basically any time an error message it's been told to log is somehow marked as 'fatal', Shell will go down this path where it winds up intentionally dying in _g_log_abort() , but my C is not exactly the greatest in the world so I may be misunderstanding something there. Anyhow, we should not be treating all crashes that go down this path as dupes, as they clearly are not.
This pull request should fix the issue: https://github.com/abrt/satyr/pull/264
*** Bug 1526166 has been marked as a duplicate of this bug. ***
Can we please get this change pushed out as an update for F26 and F27? I see the change has gone to Rawhide in a -2 package build, but it has not gone to F26 or F27 so far as I can see, and of course the majority of Fedora users are on stable releases, not Rawhide. So we are still seeing problems caused by this. The most recent one we came across is https://bugzilla.redhat.com/show_bug.cgi?id=1510059 - once again, many unrelated bugs are being closed as dupes of that (or the reporters are just getting comments added on that bug, instead of new bugs filed) due to this _g_log_abort issue. 1510059 also shows a slightly different path to _g_log_abort, perhaps one or more of these functions should also be added to this 'blacklist': #0 raise at ../sysdeps/unix/sysv/linux/raise.c:51 #1 dump_gjs_stack_on_signal_handler at ../src/main.c:372 #3 _g_log_abort at gmessages.c:554 #4 g_log_writer_default at gmessages.c:2628 #5 g_log_structured_array at gmessages.c:1941 #6 g_log_structured at gmessages.c:1768 note we get in via g_log_structured -> g_log_structured_array -> g_log_writer_default , not default_log_handler -> g_log_default_handler like all the other cases.
For the record, I found a few other bugs like this, and unpicked all those dupes too. The biggest were: https://bugzilla.redhat.com/show_bug.cgi?id=1370073 (F25) https://bugzilla.redhat.com/show_bug.cgi?id=1402492 (F26)
It's also worth noting that one fairly common case is that GNOME crashes on this codepath with the critical error message just being "Connection to xwayland lost." All that tells us, really, is that XWayland crashed; it doesn't tell us why, nor does it mean at all that all bugs which hit that codepath are dupes of each other. I'm not sure if it's possible for abrt to handle that well, but it'd be great if it could.
I've created pull request which adds g_log_structured -> g_log_structured_array -> g_log_writer_default functions into "blacklist". https://github.com/abrt/satyr/pull/270 Will backport all added "blacklisted" functions to f27 and f26.
Thanks a lot!
satyr-0.25-1.fc27 has been submitted as an update to Fedora 27. https://bodhi.fedoraproject.org/updates/FEDORA-2018-580c89505f
satyr-0.25-1.fc27 has been pushed to the Fedora 27 testing repository. If problems still persist, please make note of it in this bug report. See https://fedoraproject.org/wiki/QA:Updates_Testing for instructions on how to install test updates. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2018-580c89505f
satyr-0.25-1.fc27 has been pushed to the Fedora 27 stable repository. If problems still persist, please make note of it in this bug report.