As asked in bug https://bugs.kde.org/show_bug.cgi?id=382418, I create a new one because I confirm that on my side autosave action stop animation render. Currently on Krita 5, if autosave run while an animation is rendering, render is stopped with a popup message "failed to render animation frames”. STEPS TO REPRODUCE 1) Create an animation for which time to render is long enough to have autosave triggered 2) Ensure that document is in modified state (ie: the autosave will be triggered) 3) Render animation OBSERVED RESULTS Render is interrupted, there’s a status bar notification "autosaving postponed: document is busy" but rendering continues. A couple of seconds after, there's a new notification. This is repeated 4 times and then finally, autosave is applied. A dialog box "waiting for image operation to complete (<updating>)" is displayed for a couple of seconds and it stops the rendering with a "failed to render animation frames" message. EXPECTED RESULTS If an animation is rendering: - the autosave should properly be cancelled without stopping render or - the autosave should be temporary disabled before animation is rendered, and enabled once rendering is done (if enable before rendering) ADDITIONAL INFORMATION """How many frames Krita is trying to render (viewable on the render progress bar as [currentFrame]/[totalFrameCount].)""" This part have no importance... It occurs with 2 keyframes on my side as well as hundred of keyframes. Total number is around 600frames on my test but I think it means nothing as from a simple animation of 600 frames it could 15s to render all frame to more than one hour for a more complex 600frames animation. """How many cores your system has dedicated to rendering animations or caching frame data (should be viewable on the performance tab.)""" from 1 core to 24 Cores, the problem occurs. Last test as been made with 4 cores """Does the failure occur when saving image sequences or does it occur during the FFMPEG phase (You can verify by checking if all image frames have rendered out properly and saved to that hard disk.)""" Rendered as PNG sequence, then no ffmpeg involved in process Tested on krita-5.0.0-prealpha-dba5ce5-x86_64.appimage and many other pre-alpha version Topic on KA: https://krita-artists.org/t/krita-5-animation-curves-applied-to-transform-mask/24895/13?u=grum999 Video example: https://krita-artists.org/t/krita-5-animation-curves-applied-to-transform-mask/24895/31?u=grum999 Grum999
I actually can't reproduce this one on my system yet. In my tests so far autosaving has been paused during rendering, just like in your video. Unlike your video, the rendering eventually completes. But since you have a video of it happening I will confirm it. ----- By the way, the reason Eoin asked about the length of your animation is that when an animation contains an animated transform mask with interpolation, each individual frame is considered unique--not just the key frames. While this is not necessarily relevant, it *could be*, as bugs can get pretty weird sometimes.
(In reply to Emmet O'Neill from comment #1) > I actually can't reproduce this one on my system yet. > > In my tests so far autosaving has been paused during rendering, just like in > your video. Unlike your video, the rendering eventually completes. > > But since you have a video of it happening I will confirm it. > > ----- > > By the way, the reason Eoin asked about the length of your animation is that > when an animation contains an animated transform mask with interpolation, > each individual frame is considered unique--not just the key frames. > > While this is not necessarily relevant, it *could be*, as bugs can get > pretty weird sometimes. I'll try tomorrow to do test on my 4CPU laptop to check if I have the problem or not. I'll also try to test if I have the problem or not with Krita 4.4.3 (then doing test on document with 'hard' keyframes instead of animated curves frames) Grum999
*** This bug has been marked as a duplicate of bug 438754 ***
Git commit 00b8f331260f14792c165298857668446737d918 by Dmitry Kazakov. Committed on 05/07/2021 at 10:39. Pushed by dkazakov into branch 'master'. From https://bugs.kde.org/show_bug.cgi?id=438754#c9 I confirm this commit fixed the autosave problem Grum999