Playback used to stop at last selected frame when clicking stop or clicking on the canvas. This was very-useful expected behavior, as making a selection around a specific frame and alternate play/draw on this frame is a basic workflow action. Now it doesn't anymore, pressing stop stops at random frames, and removes the selection in the timeline! Only if I use the shortcuts for play/pause instead of clicking on the button works as expected. This is a big regression, that I would call major issue as that's really annoying while animating (and is a blocker for recording the tutorial videos..) Reproducible: Always
Ok, a few more infos: After more testing on last stable (tested both 3.0 and 3.0.1.1), looks like I could already reproduce similar issue (selection in the timeline is reset and current frame change), but only on very rare occasions. It happens mostly if I do stess-test with fast repeating, and everytime it happens I noticed the cache for the current frame is not yet loaded when I press stop. So the bug was already there, but "rarely" happening. Now it's triggered more easily making it more obvious, probably because of the cache issues in https://bugs.kde.org/show_bug.cgi?id=371528 .
Ok, now I can reproduce that easily on any version, and describe the actual issue here as: -If a cache image is computed while playing (and so I see the quick notifications ""render animation" and "updating" while playing), when I stop in any way, current frame goes to next frame, and selected area in the timeline gets reset. So that's not a regression, but still that's a bug that should be fixed.
Git commit 13bbe54e3a85cd79f5a102e6d9c7aefb70c15e16 by Dmitry Kazakov. Committed on 24/10/2016 at 12:10. Pushed by dkazakov into branch 'kazakov/undo-with-instant-preview-T2544'. Fix recalculation of the cache when editing the last frame of the layer M +2 -4 libs/ui/kis_animation_cache_populator.cpp http://commits.kde.org/krita/13bbe54e3a85cd79f5a102e6d9c7aefb70c15e16
Was this fixed or not?
This is fixed in kazakov/undo-with-instant-preview-T2544 , which also has other changes that needed testing. I've been testing it the last days, and didn't find new bugs that weren't also in master. I guess it could be merged now, though if someone else want to give some testing...