Fixed bug that caused canvas updates to trigger two compositor paints.

Fixed bug that caused canvas updates to trigger two compositor paints. LayerRendererChromium::updateLayers was triggering a scheduleComposite because the paint aggregator in RenderWidget was emptied beforehand. This change (in combination with the webkit-side patch https://bugs.webkit.org/show_bug.cgi?id=64224) delays the paint aggregator emptying until after updateLayers and before paint.

The webkit patch must land before this one.

BUG=84887
TEST=www/~jbates/test/double_frames/ should run at about 10 fps; if it's closer to 20, there is a bug

[jbates, 2011-07-09 02:06:19]

http://codereview.chromium.org/7327030/diff/1/content/renderer/render_widget.cc
File content/renderer/render_widget.cc (left):

http://codereview.chromium.org/7327030/diff/1/content/renderer/render_widget....
content/renderer/render_widget.cc:789:
HISTOGRAM_COUNTS_100("MPArch.RW_PaintRectCount", update.paint_rects.size());
None of this code block has changed - it's just moved inside the
if(!is_accelerated_compositing_active_) block

[nduca, 2011-07-12 16:54:26]

Is this patch still current in light of wkbug64224's current state?

[jbates, 2011-07-12 17:16:46]

On 2011/07/12 16:54:26, nduca wrote:
> Is this patch still current in light of wkbug64224's current state?

Probably not, no need to review for now, still testing the other side though.