Absolute-position the initial containing block.

Absolute-position the initial containing block.

LayoutView needs to clip its contents when root layer scrolling is enabled. It
already does, but only as a side effect of being composited, and some things
hit-test in front of the scrollbars because DeprecatedPaintLayerClipper thinks
absolute-positioned content can only be clipped by positioned containers.

Generally a static-positioned scroller will not clip absolute-positioned
descendants, so this seems like the cleanest fix (versus special-casing
LayoutView in DPLClipper).

I think there are other cases where the assumption in DPLClipper may be wrong
(in particular, when the scroller has a transform), but that can be addressed
separately; see http://crbug.com/499083.

Two test cases in WebFrameTest no longer pass with root layer scrolling, but
they only worked by accident. Fix tracked in http://crbug.com/499053.

BUG=495856
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=197054

[skobes, 2015-06-10 23:25:38]

Patchset #1 (id:1) has been deleted

[skobes, 2015-06-10 23:25:45]

Patchset #1 (id:20001) has been deleted

[skobes, 2015-06-10 23:25:53]

Patchset #1 (id:40001) has been deleted

[skobes, 2015-06-10 23:25:58]

Patchset #1 (id:60001) has been deleted

[skobes, 2015-06-10 23:26:03]

Patchset #1 (id:80001) has been deleted

[skobes, 2015-06-11 00:37:08]

skobes@chromium.org changed reviewers:
+ esprehn@chromium.org

[esprehn, 2015-06-12 06:16:55]

lgtm, does the spec say this somewhere?

[skobes, 2015-06-12 18:37:37]

On 2015/06/12 06:16:55, esprehn wrote:
> lgtm, does the spec say this somewhere?

As far as I can tell, CSS 2.1 does not specify that the initial containing block
is static or absolutely positioned, only that it "has the dimensions of the
viewport and is anchored at the canvas origin".

Section 11.1.1 says that overflow clip affects "all of the element's content
except any descendant elements (and their respective content and descendants)
whose containing block is the viewport or an ancestor of the element".  From
this it can be inferred that the initial containing block should clip its
absolutely positioned descendants.

[skobes, 2015-06-12 18:37:43]

The CQ bit was checked by skobes@chromium.org

[commit-bot: I haz the power, 2015-06-12 18:37:58]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1172153004/100001

[commit-bot: I haz the power, 2015-06-12 19:32:53]

Committed patchset #1 (id:100001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=197054