Description
Aron Budea
2017-01-26 02:56:40 UTC
Aron: are you sure tdf#104920 should still be related to this one? Indeed, it seems assert concerned another tracker (tdf#105077) Julien, you're right, bug 104920 doesn't cause an assertion error anymore. I'm removing the "dependency". Let's remove the non-assertions. Dávid, a good principle when evaluating if a crashing bug belongs here is: - it only happens with a build with debug symbols, - the backtrace or the crash message mentions: assertion failed. In case of regressions reproducible in Linux the daily dbgutil bibisect repos [1] could be used to bibisect a regression to the day. Some considerations: - if only the assert was introduced then, the result isn't helpful, - the offending code might be in a completely different place (the assert should give some hints in itself). [1] https://wiki.documentfoundation.org/QA/Bibisect/Linux |