FCMify text-based-repaint layout tests

FCMify text-based-repaint layout tests

Repaint tracking happens in both FrameView and GraphicsLayer.  When
layout tests are always composited, then FrameView no longer tracks
repaints and they are all collected on GraphicsLayer.  The
text-based-repaint.js test is adjusted to print out the graphics
layer tree and to force any test using it into FCM.

In order to print out repaint rects, the layer tree has to be printed
out instead.  I looked through all the results and the only differences
are that repaints are now clipped to not extend into negative space and
that repaints are relative to the graphics layer and not the viewport
(so are differently offset by scrolling).

The vast majority of these are text changes, but there are 4 image
changes in svg, that are all using "text"-based-repaint to do pixel
repaint tests.  These are all scrollbar differences similar to
https://codereview.chromium.org/249963002/ and one imperceptible
image scaling change.

A follow up to this would be to remove the FrameView-based repaint
tracking plumbing and only use the GraphicsLayer-based version.

BUG=361729
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=172545

[enne (OOO), 2014-04-23 23:56:55]

https://codereview.chromium.org/252353002/diff/20001/LayoutTests/TestExpectat...
File LayoutTests/TestExpectations (left):

https://codereview.chromium.org/252353002/diff/20001/LayoutTests/TestExpectat...
LayoutTests/TestExpectations:1042: crbug.com/362050 [ Retina ]
fast/repaint/block-layout-inline-children-replaced.html [ Failure ]
This conflicts with rebaselining on all other platforms.  I left a comment on
https://code.google.com/p/chromium/issues/detail?id=362050.  It looks like this
just needs a rebaseline.

[abarth-chromium, 2014-04-24 14:41:39]

lgtm

[abarth-chromium, 2014-04-24 14:41:53]

The CQ bit was checked by abarth@chromium.org

[commit-bot: I haz the power, 2014-04-24 14:42:13]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/enne@chromium.org/252353002/20001

[commit-bot: I haz the power, 2014-04-24 15:16:46]

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

[commit-bot: I haz the power, 2014-04-24 15:16:47]

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

[enne (OOO), 2014-04-24 18:31:43]

The Mac has a bunch of pixel differences (e.g. one bit in alpha or color) due to
force compositing mode changes.  These changes seem more correct now.  Looking
at some values, alpha=255 instead of alpha=254 for things that should be opaque
and colors match the Linux expectations better.  So, more churn, but an
improvement.

[enne (OOO), 2014-04-24 18:31:48]

The CQ bit was checked by enne@chromium.org

[commit-bot: I haz the power, 2014-04-24 18:32:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/enne@chromium.org/252353002/40001

[commit-bot: I haz the power, 2014-04-24 22:59:01]

Change committed as 172545

[enne (OOO), 2014-04-25 22:26:37]

Reverted in http://src.chromium.org/viewvc/blink?view=revision&revision=172675