tree 4a23dccf2f412cf0489acde29dfab3b539f2badd
parent 9be67046d1f0d57d0d4fba089f151ef044572f5d
author Aaron Krajeski <aaronhk@chromium.org> 1669502315 -0800
committer Blink WPT Bot <blink-w3c-test-autoroller@chromium.org> 1669503552 -0800

Make new InterpolableColor type

This makes third_party/blink/web_tests/external/wpt/css/css-color/animation/color-interpolation.html pass in its entirety. It was
not taking into account that lightness values for Oklab are in range
[0, 100] instead of [0, 1], so the expectations are adjusted here.

https://www.w3.org/TR/css-color-4/#interpolation
With the changes coming in CSS Color 4, CSS animations and transitions
need to change. Previously all interpolations happened in the sRGB
color space. Now the spec requires that new "non legacy" colors
interpolate in the Oklab colorspace while legacy colors continue to
interpolate in sRGB.

Crucially, if a legacy color is animating with or transitioning to a
non-legacy color the transition should happen in Oklab.

Previously colors that were interpolated were generic InterpolableLists.
This was fine when colors could be naively assumed to be in the same
color space, but is not no longer sufficient. This CL creates an
InterpolableColor class inheriting from InterpolableValue. It stores
the three color params generically (as opposed to "red", "green" and
"blue") along with alpha and the colorspace for interpolation. When
an interoperable pair of colors is created, they will check their
color spaces to see if they are equal, and convert their values if
necessary.

InterpolableNumber::Clone and InterpolableList::Clone were exposed as
part of this CL. It seems like they should have been all along, its
only that InterpolableColor was the first InterpolableValue type to
make use of them.

Open questions:
- CSSColorInterpolationType::ResolveInterpolableColor simply adds in
colors, this doesn't seem like it should work?
- InterpolableColor::Composite only adds in alpha if the two alpha values
are not the same. This is copy pasted from the previous code but does
not make much sense to me.

Bug: 1365635
Change-Id: I17303067890742e3c67ae9b8cdd1073b7956aea9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4024793
Reviewed-by: Robert Flack <flackr@chromium.org>
Commit-Queue: Aaron Krajeski <aaronhk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1075927}