Web Animations: Add SVGLengthListInterpolationType

Web Animations: Add SVGLengthListInterpolationType

This change migrates the following SVG attribute animations over to
SVGLengthListInterpolationType and enables additive keyframe composite modes as
well as neutral keyframes: dx, dy, y.

BUG=530436

[suzyh_UTC10 (ex-contributor), 2015-12-16 22:06:42]

suzyh@chromium.org changed reviewers:
+ alancutter@chromium.org

[fs, 2015-12-16 22:43:29]

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/LayoutTe...
File
third_party/WebKit/LayoutTests/animations/svg-attribute-composition/svg-dx-dy-composition.html
(right):

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/animations/svg-attribute-composition/svg-dx-dy-composition.html:6:
<text class="target">123</text>
Since this is testing the 'list' cases, maybe it'd make sense to somehow
indicate this in the filename? Maybe there's a convention for that already.
Anyway, only a suggestion.

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/animation/StringKeyframe.cpp (right):

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/animation/StringKeyframe.cpp:298:
applicableTypes->append(adoptPtr(new
SVGLengthListInterpolationType(attribute)));
'dx' and 'dy' can also be numbers (on <feOffset> and <feDropShadow>;
https://svgwg.org/svg2-draft/attindex.html - or the 1.1 equivalent - might be a
helpful reference for these kind of cases)

[alancutter (OOO until 2018), 2015-12-18 00:35:50]

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/animation/SVGLengthInterpolationType.h
(right):

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/animation/SVGLengthInterpolationType.h:16:
InterpolationComponent neutralInterpolableSVGLength();
Global functions like this aren't consistent with the other InterpolationTypes,
they have these kinds of functions as static member functions.
We should maintain consistency in this patch and consider altering the
convention globally later.

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/animation/SVGLengthInterpolationType.h:18:
PassRefPtrWillBeRawPtr<SVGLength> applyInterpolableSVGLength(const
InterpolableValue&, const SVGLengthContext&, const SVGLengthMode&, bool);
This doesn't actually apply anything, I'd call it appliedSVGLength() or
resolveInterpolableSVGLength().
The bool parameter needs a name, we only omit the name if it's obvious from the
type what it represents.

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/animation/SVGLengthListInterpolationType.cpp
(right):

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/animation/SVGLengthListInterpolationType.cpp:18:
conversionCheckers.append(UnderlyingLengthChecker::create(*this,
underlyingLength));
We should have a neutral keyframe responsiveness test to trigger this checker.

https://codereview.chromium.org/1535523002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/animation/SVGLengthListInterpolationType.cpp:59:
underlyingValue.set(&value);
Does SMIL simply replace like this?
We need tests for mismatched lengths.

[suzyh_UTC10 (ex-contributor), 2016-01-13 23:35:30]

suzyh@chromium.org changed reviewers:
- alancutter@chromium.org

[suzyh_UTC10 (ex-contributor), 2016-01-13 23:35:30]

NB: This patch has been subsumed by https://codereview.chromium.org/1587533002