Implicit keyframe for a CSS Animation should always use the underlying style

• Component: Animations
• Status: RESOLVED FIXED

[Antoine Quint]

Implicit keyframe for a CSS Animation should always use the underlying style

[Antoine Quint]

Created attachment 448698 [details]
Patch

[Antoine Quint]

Submitted web-platform-tests pull request: https://github.com/web-platform-tests/wpt/pull/32301

[EWS Watchlist]

This patch modifies the imported WPT tests. Please ensure that any changes on the tests (not coming from a WPT import) are exported to WPT. Please see https://trac.webkit.org/wiki/WPTExportProcess

[Antti Koivisto]

Comment on attachment 448698 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=448698&action=review

> Source/WebCore/rendering/style/KeyframeList.cpp:92
>          KeyframeValue keyframeValue(keyframe.key(), RenderStyle::clonePtr(*keyframe.style()));
>          for (auto propertyId : keyframe.properties())
>              keyframeValue.addProperty(propertyId);
> +        keyframeValue.setTimingFunction(keyframe.timingFunction());
> +        keyframeValue.setCompositeOperation(keyframe.compositeOperation());

can we just add a copy operator/function to KeyframeValue?

[Antoine Quint]

The test failure is due to bug 235019. I'll re-run EWS once that other bug lands.

[Antoine Quint]

(In reply to Antti Koivisto from comment #4)
> Comment on attachment 448698 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=448698&action=review
> 
> > Source/WebCore/rendering/style/KeyframeList.cpp:92
> >          KeyframeValue keyframeValue(keyframe.key(), RenderStyle::clonePtr(*keyframe.style()));
> >          for (auto propertyId : keyframe.properties())
> >              keyframeValue.addProperty(propertyId);
> > +        keyframeValue.setTimingFunction(keyframe.timingFunction());
> > +        keyframeValue.setCompositeOperation(keyframe.compositeOperation());
> 
> can we just add a copy operator/function to KeyframeValue?

Actually, the only two call sites for fillImplicitKeyframes() are places where we make the copies as well, so I think we can wrap this all up into a single function.

[Antoine Quint]

(In reply to Antoine Quint from comment #6)
> (In reply to Antti Koivisto from comment #4)
> > Comment on attachment 448698 [details]
> > Patch
> > 
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=448698&action=review
> > 
> > > Source/WebCore/rendering/style/KeyframeList.cpp:92
> > >          KeyframeValue keyframeValue(keyframe.key(), RenderStyle::clonePtr(*keyframe.style()));
> > >          for (auto propertyId : keyframe.properties())
> > >              keyframeValue.addProperty(propertyId);
> > > +        keyframeValue.setTimingFunction(keyframe.timingFunction());
> > > +        keyframeValue.setCompositeOperation(keyframe.compositeOperation());
> > 
> > can we just add a copy operator/function to KeyframeValue?
> 
> Actually, the only two call sites for fillImplicitKeyframes() are places
> where we make the copies as well, so I think we can wrap this all up into a
> single function.

Actually, that's not true. I'll attempt to refactor this in a separate patch.

[Antoine Quint]

Created attachment 448717 [details]
Patch for landing

[EWS]

Committed r287827 (245879@main): <https://commits.webkit.org/245879@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 448717 [details].

[Radar WebKit Bug Importer]

<rdar://problem/87314805>