Path::addBeziersForRoundedRect can emit degenerate cubics

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Tom Hudson]

Path::addBeziersForRoundedRect calls addBezierCurveTo(), even for corners where both radii are 0.
This is easy to suppress, and the resulting partially-round rects are still correct.

[Tom Hudson]

Created attachment 175451 [details]
Patch

[Stephen White]

Comment on attachment 175451 [details]
Patch

Looks fine (although this code could use some refactoring).  Please add a line to the ChangeLog about test coverage on landing.  r=me

[Tom Hudson]

Created attachment 175453 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 175453 [details]
Patch for landing

Clearing flags on attachment: 175453

Committed r135411: <http://trac.webkit.org/changeset/135411>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Jian Li]

There're quite a few test failures caused by this patch.

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Fframes%2FonlyCommentInIFrame.html%2Cfast%2Fframes%2Fvalid.html%2Cfast%2Fframes%2Fiframe-scaling-with-scroll.html%2Cfast%2Fframes%2Fiframe-scrolling-attribute.html

[Stephen White]

I'm seeing flakiness in scrollbars on Mac 10.7 Debug.

If it's just a few tests, I'd say don't roll it out just now.  My suspicion is that the patch didn't introduce new flakiness, just moved it around between tests.

[Tom Hudson]

(In reply to comment #6)
> There're quite a few test failures caused by this patch.
> 
> http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Fframes%2FonlyCommentInIFrame.html%2Cfast%2Fframes%2Fvalid.html%2Cfast%2Fframes%2Fiframe-scaling-with-scroll.html%2Cfast%2Fframes%2Fiframe-scrolling-attribute.html

I'm not sure why you're pointing at this patch; it looks to me like those have been quite flaky ever since WebKit r135309.