Created attachment 182748 [details] Composited rounded rect shows opaque black corners. PlatformContextSkia::drawRRect (added http://trac.webkit.org/changeset/139191) should call didDrawBounded, not didDrawRect.
Created attachment 182780 [details] Patch
Comment on attachment 182780 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=182780&action=review > LayoutTests/platform/chromium/TestExpectations:4219 > +# Needs platform-specific results Is there a way to do this as a reftest? Maybe the same thing, but without the translateZ, so it uses the software path?
Comment on attachment 182780 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=182780&action=review >> LayoutTests/platform/chromium/TestExpectations:4219 >> +# Needs platform-specific results > > Is there a way to do this as a reftest? Maybe the same thing, but without the translateZ, so it uses the software path? I did try a reftest first, but unfortunately the compositor seems to affect the corners' anti-aliasing.
(In reply to comment #3) > (From update of attachment 182780 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=182780&action=review > > >> LayoutTests/platform/chromium/TestExpectations:4219 > >> +# Needs platform-specific results > > > > Is there a way to do this as a reftest? Maybe the same thing, but without the translateZ, so it uses the software path? > > I did try a reftest first, but unfortunately the compositor seems to affect the corners' anti-aliasing. Ahh, of course.
Comment on attachment 182780 [details] Patch Looks good, bots willing. r=me
Comment on attachment 182780 [details] Patch Thanks, Stephen.
Comment on attachment 182780 [details] Patch Clearing flags on attachment: 182780 Committed r139745: <http://trac.webkit.org/changeset/139745>
All reviewed patches have been landed. Closing bug.