Use CSS pixels for scrolling tests.

Use DIP for scrolling tests.

DIP stands for Device-Independent Pixel. Chrome on CrOS is configured
with IsUseZoomForDSFEnabled. On high-end devices with high DPI display
the return values from __GestureCommon_GetBoundingVisibleRect will be
the number of real pixels, which causes troubles as the scrolling
coordiates are expecting DIP.

The fix here is to convert real pixel values back to DIP by dividing it
with window.devicePixelRatio.

BUG=chromium:599656
TEST=smoothness.top_25_smooth now passes on Pixel

Committed: https://chromium.googlesource.com/external/github.com/catapult-project/catapult/+/91953db7ae2b087fdc2581c39d07e3862e31fa96

[bccheng, 2016-06-14 18:07:56]

bccheng@chromium.org changed reviewers:
+ achuith@chromium.org, bccheng@chromium.org, oshima@chromium.org

[bccheng, 2016-06-14 18:09:00]

Description was changed from

==========
Use DIP for scrolling tests.

DIP stands for Device-Independent Pixel. Chrome on CrOS is configured
with IsUseZoomForDSFEnabled. On high-end devices with high DPI display
the return values from __GestureCommon_GetBoundingVisibleRect will be
the number of real pixels, which causes troubles as the scrolling
coordiates are expecting DIP.

The fix here is to convert real pixel values back to DIP by dividing it
with window.devicePixelRatio.

BUG=chromium:599656
TEST=smoothness.top_25_smooth now passes on Pixel
==========

to

==========
Use DIP for scrolling tests.

DIP stands for Device-Independent Pixel. Chrome on CrOS is configured
with IsUseZoomForDSFEnabled. On high-end devices with high DPI display
the return values from __GestureCommon_GetBoundingVisibleRect will be
the number of real pixels, which causes troubles as the scrolling
coordiates are expecting DIP.

The fix here is to convert real pixel values back to DIP by dividing it
with window.devicePixelRatio.

BUG=chromium:599656
TEST=smoothness.top_25_smooth now passes on Pixel
==========

[oshima, 2016-06-14 19:11:23]

https://codereview.chromium.org/2066573004/diff/1/telemetry/telemetry/interna...
File telemetry/telemetry/internal/actions/scroll.js (right):

https://codereview.chromium.org/2066573004/diff/1/telemetry/telemetry/interna...
telemetry/telemetry/internal/actions/scroll.js:131: rect.top + (rect.height /
window.devicePixelRatio) *
hmm, this does not look right because

* it looks to me that the rect is in css pixels, so this should be the same
value regardless of use-zoom-for-dsf is on or off.
* If the rect is scaled for some reason, you also need to scale left,top,width
as well.

Can you explain how this test is designed?

[bccheng, 2016-06-14 20:14:47]

https://codereview.chromium.org/2066573004/diff/1/telemetry/telemetry/interna...
File telemetry/telemetry/internal/actions/scroll.js (right):

https://codereview.chromium.org/2066573004/diff/1/telemetry/telemetry/interna...
telemetry/telemetry/internal/actions/scroll.js:131: rect.top + (rect.height /
window.devicePixelRatio) *
On 2016/06/14 19:11:23, oshima wrote:
> hmm, this does not look right because
> 
> * it looks to me that the rect is in css pixels, so this should be the same
> value regardless of use-zoom-for-dsf is on or off.
> * If the rect is scaled for some reason, you also need to scale left,top,width
> as well.
> 
> Can you explain how this test is designed?

With IsUseZoomForDSFEnabled the value returned to rect got doubled on high DPI
devices, where window.devicePixelRatio is 2. 

Since the test is only doing vertical scrolling so I didn't adjust the x values,
but I can do that too.

[oshima, 2016-06-14 20:18:47]

On 2016/06/14 20:14:47, bccheng wrote:
>
https://codereview.chromium.org/2066573004/diff/1/telemetry/telemetry/interna...
> File telemetry/telemetry/internal/actions/scroll.js (right):
> 
>
https://codereview.chromium.org/2066573004/diff/1/telemetry/telemetry/interna...
> telemetry/telemetry/internal/actions/scroll.js:131: rect.top + (rect.height /
> window.devicePixelRatio) *
> On 2016/06/14 19:11:23, oshima wrote:
> > hmm, this does not look right because
> > 
> > * it looks to me that the rect is in css pixels, so this should be the same
> > value regardless of use-zoom-for-dsf is on or off.
> > * If the rect is scaled for some reason, you also need to scale
left,top,width
> > as well.
> > 
> > Can you explain how this test is designed?
> 
> With IsUseZoomForDSFEnabled the value returned to rect got doubled on high DPI
> devices, where window.devicePixelRatio is 2. 

I'm wondering why the rec got doubled because css pixel shouldn't be affected by
use-zoom-for-dsf mode.

Do you know what the "this.options_.element_" is?


> 
> Since the test is only doing vertical scrolling so I didn't adjust the x
values,
> but I can do that too.

[bccheng, 2016-06-14 23:42:32]

I will spend more time to trace Chrome to find out why the rect metrics are
doubled.

[oshima, 2016-06-17 21:51:30]

ok, lgtm as a stopgap solution.

[achuithb, 2016-06-18 16:25:39]

On 2016/06/17 21:51:30, oshima wrote:
> ok, lgtm as a stopgap solution.

Owner lgtm

[bccheng, 2016-06-18 17:09:01]

The CQ bit was checked by bccheng@chromium.org

[commit-bot: I haz the power, 2016-06-18 17:09:06]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2066573004/20001

[commit-bot: I haz the power, 2016-06-18 17:30:22]

Description was changed from

==========
Use DIP for scrolling tests.

DIP stands for Device-Independent Pixel. Chrome on CrOS is configured
with IsUseZoomForDSFEnabled. On high-end devices with high DPI display
the return values from __GestureCommon_GetBoundingVisibleRect will be
the number of real pixels, which causes troubles as the scrolling
coordiates are expecting DIP.

The fix here is to convert real pixel values back to DIP by dividing it
with window.devicePixelRatio.

BUG=chromium:599656
TEST=smoothness.top_25_smooth now passes on Pixel
==========

to

==========
Use DIP for scrolling tests.

DIP stands for Device-Independent Pixel. Chrome on CrOS is configured
with IsUseZoomForDSFEnabled. On high-end devices with high DPI display
the return values from __GestureCommon_GetBoundingVisibleRect will be
the number of real pixels, which causes troubles as the scrolling
coordiates are expecting DIP.

The fix here is to convert real pixel values back to DIP by dividing it
with window.devicePixelRatio.

BUG=chromium:599656
TEST=smoothness.top_25_smooth now passes on Pixel

Committed:
https://chromium.googlesource.com/external/github.com/catapult-project/catapu...
==========

[commit-bot: I haz the power, 2016-06-18 17:30:23]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/external/github.com/catapult-project/catapu...

[bccheng, 2016-06-21 16:38:35]

A revert of this CL (patchset #2 id:20001) has been created in
https://codereview.chromium.org/2086833003/ by bccheng@chromium.org.

The reason for reverting is: Speculative revert to see if it is the cause for
crbug/621489..