There are some repaint issues in accelerated compositing mode, for example on www.ddg.gg.
Created attachment 240062 [details] Patch
Comment on attachment 240062 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=240062&action=review I think this looks fine, but I can't confirm isAcceleratedCompositing at the moment. can you confirm and then I can approve the patch? > Source/WebKit/win/WebView.cpp:6796 > + if (!isAcceleratedCompositing()) I don't have WebKit source in front of me. Can you confirm 'isAcceleratedCompositing' is identical to checking m_backingLayer for USE(CA)?
(In reply to comment #2) > Comment on attachment 240062 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=240062&action=review > Thanks for reviewing :) > I think this looks fine, but I can't confirm isAcceleratedCompositing at the > moment. can you confirm and then I can approve the patch? > > > Source/WebKit/win/WebView.cpp:6796 > > + if (!isAcceleratedCompositing()) > > I don't have WebKit source in front of me. Can you confirm > 'isAcceleratedCompositing' is identical to checking m_backingLayer for > USE(CA)? Yes, when m_backingLayer is created, m_isAcceleratedCompositing is set to true, and when m_backingLayer is set to null, m_isAcceleratedCompositing is set to false.
Comment on attachment 240062 [details] Patch This also seems to fix a drawing issue that appeared sometimes during transitions in http://www.w3.org/Talks/2012/0416-CSS-WWW2012/Demos/transitions/demo-transitions-3-shadows.html
Comment on attachment 240062 [details] Patch Clearing flags on attachment: 240062 Committed r174865: <http://trac.webkit.org/changeset/174865>
All reviewed patches have been landed. Closing bug.