Fix spurious 1px scrolling.

Fix spurious 1px scrolling.

In the case when the temporary_impl_bounds_ variable is set to
ComputeInnerViewportContainerSize() (which is device_viewport_size_ /
device_scale_factor()) the value should be rounded up, to match
ContentViewCoreImpl::GetViewportSizeDip(). The incorrect rounding
is causing the layer to be scrollable when the physical width
is not cleanly divisible by the dpi scale.

BUG=405580
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=291035

[mkosiba (inactive), 2014-08-20 16:51:32]

PTAL

[danakj, 2014-08-20 17:00:25]

https://codereview.chromium.org/496473002/diff/1/cc/layers/layer_impl.cc
File cc/layers/layer_impl.cc (right):

https://codereview.chromium.org/496473002/diff/1/cc/layers/layer_impl.cc#newc...
cc/layers/layer_impl.cc:772: // TODO(wjmaclean) Convert so that bounds returns
SizeF.
OT: A SizeF is going to be terrible for layers that are millions of pixels large
(think buildbot output), as the size of it will start to lose precision. Maybe
something else needs to be considered for this.

[aelias_OOO_until_Jul13, 2014-08-20 18:20:47]

lgtm

https://codereview.chromium.org/496473002/diff/1/cc/layers/layer_impl.cc
File cc/layers/layer_impl.cc (right):

https://codereview.chromium.org/496473002/diff/1/cc/layers/layer_impl.cc#newc...
cc/layers/layer_impl.cc:772: // TODO(wjmaclean) Convert so that bounds returns
SizeF.
On 2014/08/20 at 17:00:25, danakj wrote:
> OT: A SizeF is going to be terrible for layers that are millions of pixels
large (think buildbot output), as the size of it will start to lose precision.
Maybe something else needs to be considered for this.

The more serious issue is that layer bounds lack precision between DIPs.  This
is an underlying cause of this bug and also is a blocker for Android frontend UI
switching over to DIP coordinates.  I don't think the corner case of
8-million-pixel pages (I'm not sure even the buildbot pages are that long)
should be a greater concern, particularly since the floating point inaccuracy
would probably just cause minor rendering artifacts.  It doesn't seem worth
introducing fixed-point units just for this purpose either...

[mkosiba (inactive), 2014-08-21 09:08:05]

The CQ bit was checked by mkosiba@chromium.org

[mkosiba (inactive), 2014-08-21 09:08:08]

The CQ bit was unchecked by mkosiba@chromium.org

[mkosiba (inactive), 2014-08-21 09:08:23]

The CQ bit was checked by mkosiba@chromium.org

[commit-bot: I haz the power, 2014-08-21 09:09:43]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/mkosiba@chromium.org/496473002/1

[commit-bot: I haz the power, 2014-08-21 10:11:42]

Committed patchset #1 (1) as 291035