Make compositing mode sticky when RenderView is reused in LayoutTests.

Make compositing mode sticky when RenderView is reused in LayoutTests.

The RenderView/Widget and WebViewImpl are reused between layout tests.
But the page Settings object seems to be reset, which means the
forceCompositingMode is turned off between tests, while WebViewImpl
holds onto the m_layerTreeView pointer.

WebViewImpl::enterForceCompositingMode() already has an early out if
the mode is the same as the settings, so have RenderWidget always
call enterForceCompositingMode() so that on LayoutTests other than
the first test (where the RenderWidget is being reused) when
WebViewImpl declares it will begin compositing, we ensure that it
(re-)enters ForceCompositingMode.

This allows us to remove the extra layout() and know that the tests
will never fall out of compositing mode, so we can decide immediately
to do a composited or a software readback. Adjust the WebkitTestRunner
to match these expectations, removing its ability to fallback to
a software readback if a composited readback fails, to guarantee that
in compositing mode the readback is always working instead of hiding
failures.

R=abarth, enne, piman
BUG=370130
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=269179

[danakj, 2014-05-07 21:34:50]

I ran the layout test suite locally and get the same failures with and without
this CL.

[piman, 2014-05-07 21:38:33]

lgtm

[danakj, 2014-05-07 22:36:30]

The CQ bit was checked by danakj@chromium.org

[danakj, 2014-05-07 22:36:45]

blink bots look green CQing.

[commit-bot: I haz the power, 2014-05-07 22:37:41]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/danakj@chromium.org/272683003/40001

[danakj, 2014-05-08 15:42:37]

The CQ bit was unchecked by danakj@chromium.org

[danakj, 2014-05-08 15:43:47]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2014-05-08 15:46:42]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/danakj@chromium.org/272683003/40001

[commit-bot: I haz the power, 2014-05-09 07:06:38]

Change committed as 269179