I had noticed that if I had a delay in my animation it would get offset strangely. Further investigation makes me believe that the animation takes care of firing after the delay on its own and GraphicsLayerQt.cpp doesn't need to delay the animation's start. Is there a test case that shows its required? The attached test case seems to indicate otherwise.
Created attachment 59137 [details] Proposed patch (with testcase)
To actually use that testcase I will need bug #40841 to be applied as well as the items will remain transparent (and the timing won't be seen)
Can you make sure it works with LayoutTests/animations/fill-mode-transform.html? It's an existing test case that has delay and uses AC.
(In reply to comment #3) > Can you make sure it works with LayoutTests/animations/fill-mode-transform.html? > It's an existing test case that has delay and uses AC. Thanks for the pointer to that test, however, it fails to pass. When I take out my change it continues to fail in the same way - so I'm not thinking my change is causing the failure. Has this test passed completely before?
As long as your change doesn't create a regression wrt delays, I have no problems with it. I'll look further into that test, I think it's unstable.
(In reply to comment #5) > As long as your change doesn't create a regression wrt delays, I have no problems with it. I'll look further into that test, I think it's unstable. I'm not seeing any worse behaviour, and in the test that I included in here it makes the graphicsview behave the same as when disabled.
Created attachment 59439 [details] Rediff against trunk
*** Bug 40778 has been marked as a duplicate of this bug. ***
Comment on attachment 59439 [details] Rediff against trunk Clearing flags on attachment: 59439 Committed r62010: <http://trac.webkit.org/changeset/62010>
All reviewed patches have been landed. Closing bug.
Revision r62010 cherry-picked into qtwebkit-2.0 with commit 45fa2b8b4d6906275f948b93076c38f9539427b0