Bug 54336 – PlatformContextSkia::applyAntiAliasedClipPaths does not work for paths which have evenOdd property

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[Robin Cao]

PlatformContextSkia::applyAntiAliasedClipPaths does not work for paths which have evenOdd property.
This test (http://www.w3.org/Graphics/SVG/Test/20021112/htmlframe/full-masking-path-05-f.html) fails on Chrome.

I have a fix for this. Will upload soon.

[Robin Cao]

Created attachment 82393 [details]
patch

We need to take fill type of paths into account in applyAntiAliasedClipPaths.

[James Robinson]

I think this fine - could one of you (Mike/Stephen/Brian) sanity check it?

[James Robinson]

This is marked as a failure on win/linux in chromium currently: http://code.google.com/p/chromium/issues/detail?id=25208

[Mike Reed]

SkPath::toggleInverseFillType() is a helper that does exactly this.

[Stephen White]

+agl, since I think he wrote this code.

This looks ok to me (although we should probably use the helper fn, as Mike suggests).

[Adam Langley]

LGTM if updated with Mike's suggestion.

(Of course, I still hope that Skia gets immediate anti-aliased clipping at some point :)

Shouldn't this either include updated baselines or new layout tests? Or are they to follow in a different patch?

[James Robinson]

svg/W3C-SVG-1.1/masking-path-05-f.svg covers this test.

Robin, could you remove the line referencing this test from LayoutTests/platform/chromium/test_expectations.txt when you update this patch?  Thanks!

[James Robinson]

Comment on attachment 82393 [details]
patch

Marking this one R- since it should use the helper function.

[Robin Cao]

(In reply to comment #7)
> Robin, could you remove the line referencing this test from LayoutTests/platform/chromium/test_expectations.txt when you update this patch?  Thanks!

I can do this. One question is that i only work on linux at the moment and do i have to update baselines for chromium-win?

[James Robinson]

Oh, that is a bummer.  Just leave the test_expectations.txt line alone and we (meaning probably me) will take care of the baselines.

[Robin Cao]

Created attachment 82570 [details]
updated patch

Use the helper function to toggle fill type of paths.

[Adam Langley]

LGTM. (Note I am not a WK reviewer. You need an r+ from a real reviewer also.)

[James Robinson]

Comment on attachment 82570 [details]
updated patch

R=me. Thanks again!

[WebKit Commit Bot]

The commit-queue encountered the following flaky tests while processing attachment 82570 [details]:

http/tests/misc/uncacheable-script-repeated.html bug 51009 (author: [e-mail])
The commit-queue is continuing to process your patch.

[WebKit Commit Bot]

Comment on attachment 82570 [details]
updated patch

Clearing flags on attachment: 82570

Committed r78751: <http://trac.webkit.org/changeset/78751>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.