https://bugs.kde.org/show_bug.cgi?id=444516
--- Comment #21 from Dmitry Kazakov <dimul...@gmail.com> --- Hm.. the crash logs look really weird. Here is what I can tell from inspecting them: 1) "Still unsignalled after processed 129 tiles" should not be related to the crash. It is just a debug notification, we should probably hide it from the release. 2) The crash happens in the GUI thread, most probably, due to a memory corruption (see fluctuations between SIGSEGV and SIGBUS) 3) This memory corruption happens long before the actual crash 4) In at least 4 crashes of 12, the worker tread has been working on ending the stroke (destroying the brushes and superseded undo commands). That might be related actually. I will make a build with ASAN, so David could catch then actual memory corruption :) -- You are receiving this mail because: You are watching all bug changes.
