Remove compositing trigger for will-change: top (and similar)

Remove compositing trigger for will-change: top (and similar)

We can't use will-change: top (and similar) as a compositing trigger because it
changes the paint order by giving the RenderObject a RenderLayer of type
NormalLayer. That will cause us to paint incorrectly.

In a future CL, we could consider this property as a compositing hint if the
RenderObject already has a self-painting RenderLayer. However, let's first
restore correctness in the system before making that change.

R=esprehn@chromium.org
BUG=370604
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=173574

[ajuma, 2014-05-07 20:16:43]

https://codereview.chromium.org/271753002/diff/20001/Source/core/rendering/st...
File Source/core/rendering/style/RenderStyle.cpp (left):

https://codereview.chromium.org/271753002/diff/20001/Source/core/rendering/st...
Source/core/rendering/style/RenderStyle.cpp:876: case CSSPropertyWebkitFilter:
Are these removals still needed for correctness given the changes made to layer
creation logic in RenderBox and RenderTableRow? Since we now never specifically
create RenderLayers for will-change hints, there seems to be no harm leaving
these in (since they'll only cause already-existing layers to composite).

[abarth-chromium, 2014-05-07 20:19:59]

https://codereview.chromium.org/271753002/diff/20001/Source/core/rendering/st...
File Source/core/rendering/style/RenderStyle.cpp (left):

https://codereview.chromium.org/271753002/diff/20001/Source/core/rendering/st...
Source/core/rendering/style/RenderStyle.cpp:876: case CSSPropertyWebkitFilter:
On 2014/05/07 20:16:43, ajuma wrote:
> Are these removals still needed for correctness given the changes made to
layer
> creation logic in RenderBox and RenderTableRow? Since we now never
specifically
> create RenderLayers for will-change hints, there seems to be no harm leaving
> these in (since they'll only cause already-existing layers to composite).

As I wrote in the change description, I think we should restore the system to
correctness and then possibly re-introduce these compositing triggers in the
future.  Currently they don't work.

[ajuma, 2014-05-07 21:01:03]

On 2014/05/07 20:19:59, abarth wrote:
>
https://codereview.chromium.org/271753002/diff/20001/Source/core/rendering/st...
> File Source/core/rendering/style/RenderStyle.cpp (left):
> 
>
https://codereview.chromium.org/271753002/diff/20001/Source/core/rendering/st...
> Source/core/rendering/style/RenderStyle.cpp:876: case CSSPropertyWebkitFilter:
> On 2014/05/07 20:16:43, ajuma wrote:
> > Are these removals still needed for correctness given the changes made to
> layer
> > creation logic in RenderBox and RenderTableRow? Since we now never
> specifically
> > create RenderLayers for will-change hints, there seems to be no harm leaving
> > these in (since they'll only cause already-existing layers to composite).
> 
> As I wrote in the change description, I think we should restore the system to
> correctness and then possibly re-introduce these compositing triggers in the
> future.  Currently they don't work.

lgtm

[abarth-chromium, 2014-05-07 21:01:15]

The CQ bit was checked by abarth@chromium.org

[abarth-chromium, 2014-05-07 21:01:21]

Thanks!

[commit-bot: I haz the power, 2014-05-07 21:02:10]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/abarth@chromium.org/271753002/20001

[commit-bot: I haz the power, 2014-05-07 22:16:10]

FYI, CQ is re-trying this CL (attempt #1).
Please consider checking whether the failures are real,
and report flakes to chrome-troopers@google.com.
The failing builders are:
  win_blink_rel on tryserver.blink

[commit-bot: I haz the power, 2014-05-07 23:14:49]

Change committed as 173574