Added experimental Touch.tilt, Touch.tiltDirection support for Android

For a stylus, Touch.tilt reports the angle of the stylus in degrees, in range [0, 90). Touch.tiltDirection is an clockwise angle [-180..180) in relation to the 0X axis.
The angle indicates the direction in which the stylus is tilted. 

Currently available behind the enable-experimental-web-platform-features flag.

Draft spec: https://docs.google.com/document/d/1hRjHN9T3RBKJV6LL6WQMEaw35YD_T-G41X-ZciW18xU/edit

A link to the Blink patch (https://codereview.chromium.org/750013004/)

BUG=393462

[d.pikalov, 2014-11-26 16:10:33]

d.pikalov@partner.samsung.com changed reviewers:
+ jdduke@chromium.org, raymes@chromium.org

[jdduke (slow), 2014-11-26 18:07:26]

On 2014/11/26 16:10:33, d.pikalov wrote:
> mailto:d.pikalov@partner.samsung.com changed reviewers:
> + mailto:jdduke@chromium.org, mailto:raymes@chromium.org

There's no bug for this? This looks like we're just funneling more data that
ends up going nowhere. Do you have a proposal to change the web-facing touch
event spec?

[jdduke (slow), 2014-11-26 18:13:29]

Please include in the description:

A link to the Blink patch (https://codereview.chromium.org/750013004/)
A link to the BUG that has more details about the change.
A few more details about the path toward formalizing this as a web-facing
property.

https://codereview.chromium.org/755403006/diff/1/ui/events/gesture_detection/...
File ui/events/gesture_detection/motion_event.h (right):

https://codereview.chromium.org/755403006/diff/1/ui/events/gesture_detection/...
ui/events/gesture_detection/motion_event.h:98: float GetTilt() const { return
GetTilt(0); }
This property probably deserves a comment as the Android description is
non-trivial
(http://developer.android.com/reference/android/view/MotionEvent.html#AXIS_ORI...).

[d.pikalov, 2014-11-28 13:27:45]

On 2014/11/26 18:07:26, jdduke wrote:
> On 2014/11/26 16:10:33, d.pikalov wrote:
> > mailto:d.pikalov@partner.samsung.com changed reviewers:
> > + mailto:jdduke@chromium.org, mailto:raymes@chromium.org
> 
> There's no bug for this? This looks like we're just funneling more data that
> ends up going nowhere. Do you have a proposal to change the web-facing touch
> event spec?

Thanks for comments. Yes, there is no bug for this, yet.

I'll open CR with detailed proposal asap.

[d.pikalov, 2014-11-28 13:55:13]

https://codereview.chromium.org/755403006/diff/1/ui/events/gesture_detection/...
File ui/events/gesture_detection/motion_event.h (right):

https://codereview.chromium.org/755403006/diff/1/ui/events/gesture_detection/...
ui/events/gesture_detection/motion_event.h:98: float GetTilt() const { return
GetTilt(0); }
On 2014/11/26 18:13:29, jdduke wrote:
> This property probably deserves a comment as the Android description is
> non-trivial
>
(http://developer.android.com/reference/android/view/MotionEvent.html#AXIS_ORI...).

Done.

[d.pikalov, 2015-02-17 09:59:37]

d.pikalov@partner.samsung.com changed reviewers:
+ sadrul@chromium.org

[jdduke (slow), 2015-02-20 16:20:13]

https://codereview.chromium.org/755403006/diff/100001/ui/events/gesture_detec...
File ui/events/gesture_detection/motion_event.h (right):

https://codereview.chromium.org/755403006/diff/100001/ui/events/gesture_detec...
ui/events/gesture_detection/motion_event.h:63: virtual float GetTilt(size_t
pointer_index) const = 0;
Interesting, it looks like Android repurposes orientation for stylus inputs to
convey tilt:
http://developer.android.com/reference/android/view/MotionEvent.html#AXIS_ORI....

[Rick Byers, 2015-02-20 21:19:07]

rbyers@chromium.org changed reviewers:
+ rbyers@chromium.org

[Rick Byers, 2015-02-20 21:19:08]

https://codereview.chromium.org/755403006/diff/100001/ui/events/gesture_detec...
File ui/events/gesture_detection/motion_event.h (right):

https://codereview.chromium.org/755403006/diff/100001/ui/events/gesture_detec...
ui/events/gesture_detection/motion_event.h:63: virtual float GetTilt(size_t
pointer_index) const = 0;
On 2015/02/20 16:20:13, jdduke wrote:
> Interesting, it looks like Android repurposes orientation for stylus inputs to
> convey tilt:
>
http://developer.android.com/reference/android/view/MotionEvent.html#AXIS_ORI....

That's TiltDirection, yes.  I tried asking them if they felt conveying touch
orientation and stylus rotation with the same API (with different documentation
/ range of values for the two cases) was a good thing or a bad thing but they
just said they had no data.  Since the ranges (and likely use cases) are
different, I personally think using a separate API like this is better.  I
believe we discussed this on public-touchevents.

[Rick Byers, 2015-03-05 15:01:02]

Reviewed event_sender changes.

https://codereview.chromium.org/755403006/diff/120001/content/shell/renderer/...
File content/shell/renderer/test_runner/event_sender.cc (right):

https://codereview.chromium.org/755403006/diff/120001/content/shell/renderer/...
content/shell/renderer/test_runner/event_sender.cc:1567: if
(!args->PeekNext().IsEmpty()) {
Is there a good reason to allow tilt to be set two different ways (via
updateTouchPoint and via setTouchTilt)?  It seems odd that we have this for tilt
but not radius.  Unless there's a good reason, I'd prefer we just have
setTouchPointTilt.

Ironically it looks like some tests are designed to support a
setTouchPointRadius method (basic-single-touch-events.js), but we don't actually
have such a method.  If you're interested, that would be a great thing to fix at
the same time as this (in another CL).  Ideally the pattern for radius and tilt
should be the same.

[d.pikalov, 2015-03-09 13:39:50]

https://codereview.chromium.org/755403006/diff/120001/content/shell/renderer/...
File content/shell/renderer/test_runner/event_sender.cc (right):

https://codereview.chromium.org/755403006/diff/120001/content/shell/renderer/...
content/shell/renderer/test_runner/event_sender.cc:1567: if
(!args->PeekNext().IsEmpty()) {
On 2015/03/05 15:01:02, Rick Byers wrote:
> Is there a good reason to allow tilt to be set two different ways (via
> updateTouchPoint and via setTouchTilt)?  It seems odd that we have this for
tilt
> but not radius.  Unless there's a good reason, I'd prefer we just have
> setTouchPointTilt.
Done.

> 
> Ironically it looks like some tests are designed to support a
> setTouchPointRadius method (basic-single-touch-events.js), but we don't
actually
> have such a method.  If you're interested, that would be a great thing to fix
at
> the same time as this (in another CL).  Ideally the pattern for radius and
tilt
> should be the same. 

Please review
https://codereview.chromium.org/990193002
https://codereview.chromium.org/990183003

[d.pikalov, 2015-03-17 10:51:39]

Added parameter bool 'set_state_moved' to eventSender.SetTouchPointTilt()

[Rick Byers, 2015-03-17 14:40:35]

https://codereview.chromium.org/755403006/diff/160001/content/shell/renderer/...
File content/shell/renderer/test_runner/event_sender.cc (right):

https://codereview.chromium.org/755403006/diff/160001/content/shell/renderer/...
content/shell/renderer/test_runner/event_sender.cc:1821: if
(!args->PeekNext().IsEmpty()) {
you can remove this chunk now, right?

[d.pikalov, 2015-03-17 16:15:07]

Done

https://codereview.chromium.org/755403006/diff/160001/content/shell/renderer/...
File content/shell/renderer/test_runner/event_sender.cc (right):

https://codereview.chromium.org/755403006/diff/160001/content/shell/renderer/...
content/shell/renderer/test_runner/event_sender.cc:1821: if
(!args->PeekNext().IsEmpty()) {
On 2015/03/17 14:40:35, Rick Byers wrote:
> you can remove this chunk now, right?

Done. I'll remove chunk for radius in CL 990183003