Fix RenderGeometryMap ASSERT in touch hit rect computation

Fix RenderGeometryMap ASSERT in touch hit rect computation

RenderGeometryMap requires that the RenderView be pushed as the first mapping.
It helpfully does this for you when on the 'canMapBetweenRenderers' fast-path,
but not in the slow path.  Touch hit rect computation wasn't explicitly
pushing the RenderView and so could trigger this ASSERT (and potentially
other issues) if the first layer pushed doesn't satisfy the canMapBetweenRenderers
requirement (almost never the case in practice since the first layer is the
document element's layer).

BUG=339141
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=166094

[Rick Byers, 2014-01-30 03:09:22]

Levi,
Please review this tiny hit-test computation tweak.

Thanks,
   Rick

[leviw_travelin_and_unemployed, 2014-01-30 03:13:01]

Makes sense! LGTM.

https://codereview.chromium.org/129813009/diff/1/LayoutTests/fast/events/touc...
File LayoutTests/fast/events/touch/touch-rect-assert-first-layer-special.html
(right):

https://codereview.chromium.org/129813009/diff/1/LayoutTests/fast/events/touc...
LayoutTests/fast/events/touch/touch-rect-assert-first-layer-special.html:17:
<div id="touchtarget" ontouchstart="alert()">Foo</div>
Why the alert? If it's for manual testing, you may want to mention it in the
description.

https://codereview.chromium.org/129813009/diff/1/LayoutTests/fast/events/touc...
LayoutTests/fast/events/touch/touch-rect-assert-first-layer-special.html:21: if
(window.internals) {
Nit: braces for one-line if statement.

[Rick Byers, 2014-01-30 03:46:38]

On 2014/01/30 03:13:01, Levi wrote:
> Makes sense! LGTM.

Thanks!

>
https://codereview.chromium.org/129813009/diff/1/LayoutTests/fast/events/touc...
> File LayoutTests/fast/events/touch/touch-rect-assert-first-layer-special.html
> (right):
> 
>
https://codereview.chromium.org/129813009/diff/1/LayoutTests/fast/events/touc...
> LayoutTests/fast/events/touch/touch-rect-assert-first-layer-special.html:17:
> <div id="touchtarget" ontouchstart="alert()">Foo</div>
> Why the alert? If it's for manual testing, you may want to mention it in the
> description.

Nah, this is just a dummy expression - it's irrelevant.  I changed it to a
non-existent symbol to make that clear.
 
>
https://codereview.chromium.org/129813009/diff/1/LayoutTests/fast/events/touc...
> LayoutTests/fast/events/touch/touch-rect-assert-first-layer-special.html:21:
if
> (window.internals) {
> Nit: braces for one-line if statement.

Done

[commit-bot: I haz the power, 2014-01-30 03:49:11]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rbyers@chromium.org/129813009/20001

[commit-bot: I haz the power, 2014-01-30 06:31:04]

Change committed as 166094

[commit-bot: I haz the power, 2014-01-30 06:33:40]

CQ bit was unchecked on CL. Ignoring.