https://bugs.kde.org/show_bug.cgi?id=406950
--- Comment #8 from ella...@eskimo.com --- (In reply to Jasem Mutlaq from comment #6) > How about running just the release build (without valgrind). Any stability > issues encountered?? It's difficult for me to quantify that: the nature of memory corruption is that you experience uncorrelated crashes. More importantly: I just started using kstars so I can't compare to old versions (which apparently had fewer threads?), but for me 3.x is extremely unstable. I'm not sure I've ever had a session which did not crash. For example, I attempted to repro Bug #406638 with this bug fix on my Raspberry pi to see if it was related, and I just gave up for now. indi_simulator_guide crashes 100% of the time when I run through my repro sequence on ARM today (it crashes when I capture in the Guide tab). I also experienced kstars crashing but wasn't able to get any debug info from the core. I also saw at least one crash with this patch on amd64 but didn't acquire any usable debug info there either. I think the patch improves things, at least. (In reply to Jasem Mutlaq from comment #7) > does not exist cleanly. I haven't found what leads to this situation > sometimes, it's rare but happens. Have I got a nice solution for you! Run it under valgrind, all sorts of fun race conditions appear. :-) More generally, I'd suggest adding random time jitter debugging to make these sorts of bugs appear more often. For example, build with a random sleep around lock/unlock calls and then test. -- You are receiving this mail because: You are watching all bug changes.
