Separate the geometry update from rebuilding the GraphicsLayer tree

Separate the geometry update from rebuilding the GraphicsLayer tree

This CL separates the work we do to update the GraphicsLayer geometry from the
work we do to rebuild the GraphicsLayer tree. We now update the geometry in
tree-order rather than paint order, which fixes a bug in how we're propagating
dirty bits during the tree walk. The bug fix is captured by the
reorder-z-with-style.html test added in this CL.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=170277

[esprehn, 2014-03-28 02:24:08]

lgtm

https://codereview.chromium.org/215063006/diff/1/LayoutTests/compositing/reor...
File LayoutTests/compositing/reorder-z-with-style.html (right):

https://codereview.chromium.org/215063006/diff/1/LayoutTests/compositing/reor...
LayoutTests/compositing/reorder-z-with-style.html:43:
document.getElementById('parent').style.WebkitTransformStyle = 'flat';
The property is actually lowercase webkitTransformStyle. You're relying on a
quirk in how we process the properties.

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
File Source/core/rendering/compositing/GraphicsLayerUpdater.cpp (right):

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
Source/core/rendering/compositing/GraphicsLayerUpdater.cpp:150: for
(RenderLayer* child = layer.firstChild(); child; child = child->nextSibling())
We should add a travserseNext() like method or a Traversal:: API like the DOM
and make this process iterative. Then we can skip all the recursion overhead.

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
File Source/core/rendering/compositing/RenderLayerCompositor.cpp (right):

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
Source/core/rendering/compositing/RenderLayerCompositor.cpp:526:
TRACE_EVENT0("blink_rendering", "GraphicsLayerUpdater::updateRecursive");
The ones in recalcstyle are all webkit, should we change them to
blink_rendering?

[ojan, 2014-03-28 02:26:16]

lgtm

https://codereview.chromium.org/215063006/diff/1/LayoutTests/compositing/reor...
File LayoutTests/compositing/reorder-z-with-style.html (right):

https://codereview.chromium.org/215063006/diff/1/LayoutTests/compositing/reor...
LayoutTests/compositing/reorder-z-with-style.html:39:
testRunner.waitUntilDone();
Nit: inconsistent indentation!

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
File Source/core/rendering/compositing/RenderLayerCompositor.cpp (right):

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
Source/core/rendering/compositing/RenderLayerCompositor.cpp:525: // most of the
time, geometry is updated via RenderLayer::styleChanged().
I know this isn't new to this patch, but this comment is confusing to me.

[commit-bot: I haz the power, 2014-03-28 02:26:24]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/215063006/1

[abarth-chromium, 2014-03-28 02:35:32]

The CQ bit was unchecked by abarth@chromium.org

[abarth-chromium, 2014-03-28 02:37:14]

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
File Source/core/rendering/compositing/GraphicsLayerUpdater.cpp (right):

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
Source/core/rendering/compositing/GraphicsLayerUpdater.cpp:150: for
(RenderLayer* child = layer.firstChild(); child; child = child->nextSibling())
On 2014/03/28 02:24:09, esprehn wrote:
> We should add a travserseNext() like method or a Traversal:: API like the DOM
> and make this process iterative. Then we can skip all the recursion overhead.

+1

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
File Source/core/rendering/compositing/RenderLayerCompositor.cpp (right):

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
Source/core/rendering/compositing/RenderLayerCompositor.cpp:525: // most of the
time, geometry is updated via RenderLayer::styleChanged().
On 2014/03/28 02:26:16, ojan wrote:
> I know this isn't new to this patch, but this comment is confusing to me.

It's more wrong now.  I'll remove it.

https://codereview.chromium.org/215063006/diff/1/Source/core/rendering/compos...
Source/core/rendering/compositing/RenderLayerCompositor.cpp:526:
TRACE_EVENT0("blink_rendering", "GraphicsLayerUpdater::updateRecursive");
On 2014/03/28 02:24:09, esprehn wrote:
> The ones in recalcstyle are all webkit, should we change them to
> blink_rendering?

Dunno.  Probably?  "webkit" is so 2012

[abarth-chromium, 2014-03-28 03:47:34]

The CQ bit was checked by abarth@chromium.org

[commit-bot: I haz the power, 2014-03-28 03:47:37]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/215063006/20001

[commit-bot: I haz the power, 2014-03-28 03:54:14]

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

[commit-bot: I haz the power, 2014-03-28 03:54:15]

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

[abarth-chromium, 2014-03-28 03:56:02]

The CQ bit was checked by abarth@chromium.org

[commit-bot: I haz the power, 2014-03-28 03:56:29]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/215063006/20001

[commit-bot: I haz the power, 2014-03-28 04:53:02]

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

[commit-bot: I haz the power, 2014-03-28 04:53:03]

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

[abarth-chromium, 2014-03-28 04:55:21]

The CQ bit was checked by abarth@chromium.org

[commit-bot: I haz the power, 2014-03-28 04:55:30]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/abarth@chromium.org/215063006/20001

[commit-bot: I haz the power, 2014-03-28 06:02:06]

Change committed as 170277