Resolve style in element.animate()

Resolve style in element.animate()

This reverts the change to EffectInput::convert() made by
https://codereview.chromium.org/343563002.

This fixes issues when computing compositor animation values using an
out of date element style.

BUG=495922
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=196479

[alancutter (OOO until 2018), 2015-06-04 04:46:13]

alancutter@chromium.org changed reviewers:
+ dstockwell@chromium.org, esprehn@chromium.org

[alancutter (OOO until 2018), 2015-06-04 04:46:14]

Confirmed locally that this does not regress performance on Nexus 10 composited
tough_animation_cases.

[esprehn, 2015-06-04 06:09:13]

Looks like you made it fail lots of unit tests, now it doesn't throw an
exception which is weird?

[shans, 2015-06-04 06:15:31]

On 2015/06/04 06:09:13, esprehn wrote:
> Looks like you made it fail lots of unit tests, now it doesn't throw an
> exception which is weird?

Looks like Alan just missed a few tests.

[esprehn, 2015-06-04 06:18:10]

https://codereview.chromium.org/1153413007/diff/20001/Source/core/animation/E...
File Source/core/animation/EffectInput.cpp (right):

https://codereview.chromium.org/1153413007/diff/20001/Source/core/animation/E...
Source/core/animation/EffectInput.cpp:199: if (!element->inActiveDocument())
This unsafe thing just for unit tests seems bad, the unit tests should put the
element into an active document. Indeed even
StyleResolver::createAnimatableValueSnapshot tries to use the document, and
those methods shouldn't be static...

[esprehn, 2015-06-04 06:20:30]

On 2015/06/04 at 06:18:10, esprehn wrote:
>
https://codereview.chromium.org/1153413007/diff/20001/Source/core/animation/E...
> File Source/core/animation/EffectInput.cpp (right):
> 
>
https://codereview.chromium.org/1153413007/diff/20001/Source/core/animation/E...
> Source/core/animation/EffectInput.cpp:199: if (!element->inActiveDocument())
> This unsafe thing just for unit tests seems bad, the unit tests should put the
element into an active document. Indeed even
StyleResolver::createAnimatableValueSnapshot tries to use the document, and
those methods shouldn't be static...

I'm actually surprised the unit tests don't crash,
StyleResolverState::StyleResolverState does 

   
m_elementStyleResources.setDeviceScaleFactor(document.frameHost()->deviceScaleFactor());

and only active documents have a frameHost(). I guess the unit tests don't test
the path where we resolve composited values?

[alancutter (OOO until 2018), 2015-06-04 06:33:11]

On 2015/06/04 at 06:20:30, esprehn wrote:
> On 2015/06/04 at 06:18:10, esprehn wrote:
> >
https://codereview.chromium.org/1153413007/diff/20001/Source/core/animation/E...
> > File Source/core/animation/EffectInput.cpp (right):
> > 
> >
https://codereview.chromium.org/1153413007/diff/20001/Source/core/animation/E...
> > Source/core/animation/EffectInput.cpp:199: if (!element->inActiveDocument())
> > This unsafe thing just for unit tests seems bad, the unit tests should put
the element into an active document. Indeed even
StyleResolver::createAnimatableValueSnapshot tries to use the document, and
those methods shouldn't be static...
> 
> I'm actually surprised the unit tests don't crash,
StyleResolverState::StyleResolverState does 
> 
>    
m_elementStyleResources.setDeviceScaleFactor(document.frameHost()->deviceScaleFactor());
> 
> and only active documents have a frameHost(). I guess the unit tests don't
test the path where we resolve composited values?

Correct, they're animating width/height that don't get composited and never
attempt to snapshot AnimatableValues.
Added DummyPageHolder and removed unsafe parameter, thanks for the tip!

[esprehn, 2015-06-04 06:35:01]

lgtm

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
File LayoutTests/animations/composited-animation-style-update-expected.html
(right):

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
LayoutTests/animations/composited-animation-style-update-expected.html:1:
<style>
ditto.

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
File LayoutTests/animations/composited-animation-style-update.html (right):

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
LayoutTests/animations/composited-animation-style-update.html:1: <style>
<!DOCTYPE html>

[alancutter (OOO until 2018), 2015-06-04 06:45:56]

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
File LayoutTests/animations/composited-animation-style-update-expected.html
(right):

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
LayoutTests/animations/composited-animation-style-update-expected.html:1:
<style>
On 2015/06/04 at 06:35:01, esprehn wrote:
> ditto.

Done.

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
File LayoutTests/animations/composited-animation-style-update.html (right):

https://codereview.chromium.org/1153413007/diff/40001/LayoutTests/animations/...
LayoutTests/animations/composited-animation-style-update.html:1: <style>
On 2015/06/04 at 06:35:01, esprehn wrote:
> <!DOCTYPE html>

Done.

[alancutter (OOO until 2018), 2015-06-04 06:47:04]

The CQ bit was checked by alancutter@chromium.org

[alancutter (OOO until 2018), 2015-06-04 06:47:05]

The patchset sent to the CQ was uploaded after l-g-t-m from esprehn@chromium.org
Link to the patchset: https://codereview.chromium.org/1153413007/#ps60001
(title: "Add DOCTYPE")

[commit-bot: I haz the power, 2015-06-04 06:47:59]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1153413007/60001

[commit-bot: I haz the power, 2015-06-04 08:24:27]

Committed patchset #4 (id:60001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=196479