Compute style-determined compositing reasons during recalc style

Compute style-determined compositing reasons during recalc style

There's no need to constantly recompute this data. It's
already determined by the style information, so we might as
well just compute it once during style recalc.

Note: We're still computing all the direct compositing
reasons during style recalc (via updateLayerCompositingState in RenderLayer::styleChanged), but now we skip recomputing
the style determined reasons again when updating layout
and compositing. ASSERTs verify that we get the same answer
as before.

This CL reduces the amount of time spent in the compositing
update for dragging the drawer in the Polymer calculator by
3%.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=169766

[abarth-chromium, 2014-03-21 01:39:47]

Not quite ready for review yet, but getting closer.  :)

[abarth-chromium, 2014-03-21 01:52:32]

This CL deletes about 3% of the profile on the calculator.  I'll clean it up and
get it ready for landing tomorrow.

[abarth-chromium, 2014-03-21 14:12:50]

This CL is ready to review now.

[ojan, 2014-03-21 17:12:06]

lgtm

https://codereview.chromium.org/207513002/diff/20001/LayoutTests/compositing/...
File LayoutTests/compositing/will-change/gpu-rasterization-hint.html (right):

https://codereview.chromium.org/207513002/diff/20001/LayoutTests/compositing/...
LayoutTests/compositing/will-change/gpu-rasterization-hint.html:84: // Force an
update of the compositing triggers.
Perhaps setAcceleratedCompositingForGpuRasterizationHintEnabled should do this
automatically?

https://codereview.chromium.org/207513002/diff/20001/Source/core/rendering/Re...
File Source/core/rendering/RenderLayer.h (right):

https://codereview.chromium.org/207513002/diff/20001/Source/core/rendering/Re...
Source/core/rendering/RenderLayer.h:629: friend class CompositingReasonFinder;
Instead of adding to the friend madness in RenderLayer, could we just make the
right things off CompositingReasonFinder public?

[abarth-chromium, 2014-03-21 17:16:13]

https://codereview.chromium.org/207513002/diff/20001/LayoutTests/compositing/...
File LayoutTests/compositing/will-change/gpu-rasterization-hint.html (right):

https://codereview.chromium.org/207513002/diff/20001/LayoutTests/compositing/...
LayoutTests/compositing/will-change/gpu-rasterization-hint.html:84: // Force an
update of the compositing triggers.
On 2014/03/21 17:12:06, ojan wrote:
> Perhaps setAcceleratedCompositingForGpuRasterizationHintEnabled should do this
> automatically?

I looked at that, but it's autogenerated by Settings.in.  In production, these
flags are just set once when the WebView is created.

https://codereview.chromium.org/207513002/diff/20001/Source/core/rendering/Re...
File Source/core/rendering/RenderLayer.h (right):

https://codereview.chromium.org/207513002/diff/20001/Source/core/rendering/Re...
Source/core/rendering/RenderLayer.h:629: friend class CompositingReasonFinder;
On 2014/03/21 17:12:06, ojan wrote:
> Instead of adding to the friend madness in RenderLayer, could we just make the
> right things off CompositingReasonFinder public?

Sure.  Will do.

[abarth-chromium, 2014-03-21 17:43:21]

The CQ bit was checked by abarth@chromium.org

[commit-bot: I haz the power, 2014-03-21 17:43:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/207513002/40001

[commit-bot: I haz the power, 2014-03-21 18:47:28]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-03-21 18:47:29]

Try jobs failed on following builders:
  tryserver.blink on win_blink_rel

[esprehn, 2014-03-21 19:01:35]

The CQ bit was checked by esprehn@chromium.org

[commit-bot: I haz the power, 2014-03-21 19:01:40]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/207513002/40001

[commit-bot: I haz the power, 2014-03-21 19:49:34]

Change committed as 169766