[Webkit-unassigned] [Bug 202739] [GTK] Implement conic gradient

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=202739

--- Comment #5 from Diego Pino <dpino at igalia.com> ---
I didn't read the previous comment until I pushed a new patch. I will change the name of the functions in a new patch.

I have fixed most of the failing tests. There were two kind of issues:

- Wrong positioning of some sectors due to making all the sectors of the circle share a common center point. It's necessary to compute a different center for each sector depending on its position in the circle.

- Some gradient colors do not match. This might be due to how Cairo computes gradient colors or might be a problem in the current code. When a conic gradient has only two colors, I need to compute a middle stop color because it's not possible to paint a whole circle with a single Bezier line. Perhaps the middle color I compute is not the same as a graphics library would compute, thus the difference. In any case, the difference is minimal (0.01%) and I think the test should be marked as passed somehow. I manage to create a new expected tests for one of these failing tests, but the other two are non trivial so I added the tests to GTK/TestExpectations as ImageFailure.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20191011/2fb1e971/attachment.html>