This causes the failures in these tests: imported/blink/fast/canvas/canvas-clip-stack-persistence.html [ ImageOnlyFailure ] imported/blink/fast/canvas/canvas-state-persistence-no-dirty.html [ ImageOnlyFailure ] imported/blink/fast/canvas/canvas-unbalanced-save.html [ ImageOnlyFailure ]
Created attachment 421382 [details] Patch
Created attachment 421428 [details] Patch
<rdar://problem/74964300>
Comment on attachment 421428 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=421428&action=review > Source/WebCore/ChangeLog:8 > + This is not the ultimate solution. But flushDrawingContext() has a bug What is the flushDrawingContext() bug?
Comment on attachment 421428 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=421428&action=review >> Source/WebCore/ChangeLog:8 >> + This is not the ultimate solution. But flushDrawingContext() has a bug > > What is the flushDrawingContext() bug? I was wrong about flushDrawingContext(). I thought there is a synchronization bug with replaying back the DisplayList items in GPUP and getting the NativeImage of the IOSurface in WebP. It turned out it is a 2D canvas compositing issue which was fixed by r273956
Created attachment 422475 [details] Patch
I am going to r+ and cq+ this patch since it is just test gardening only.
Committed r274032: <https://commits.webkit.org/r274032> All reviewed patches have been landed. Closing bug and clearing flags on attachment 422475 [details].