Add maxTouchPoints for pointer events

I need to move this issue to the email address that I used on the CLA, so closing this one and opening a new one. >_<

Replacement: https://codereview.chromium.org/27750002

[Rick Byers, 2013-08-07 13:46:51]

https://codereview.chromium.org/20598008/diff/1/Source/modules/modules.gypi
File Source/modules/modules.gypi (right):

https://codereview.chromium.org/20598008/diff/1/Source/modules/modules.gypi#n...
Source/modules/modules.gypi:182: 'pointerevents/NavigatorPointerEvents.idl',
I'm not sure if a 'module' is really appropriate for pointer events or not...

https://codereview.chromium.org/20598008/diff/1/Source/modules/pointerevents/...
File Source/modules/pointerevents/NavigatorPointerEvents.cpp (right):

https://codereview.chromium.org/20598008/diff/1/Source/modules/pointerevents/...
Source/modules/pointerevents/NavigatorPointerEvents.cpp:68: return
NavigatorPointerEvents::from(navigator)->pointerEnabled();
As discussed, I wouldn't put pointerEnabled in here yet - better to focus just
on maxTouchPoints.

https://codereview.chromium.org/20598008/diff/1/Source/modules/pointerevents/...
Source/modules/pointerevents/NavigatorPointerEvents.cpp:84: return
GetSystemMetrics(SM_MAXIMUMTOUCHES);
I suspect this won't work in the sandbox, and it really should be provided by
the browser anyway.  You probably want to plumb it through WebSettings like
device_supports_touch.  Alternately, perhaps we should better handle touch
screens coming and going during the life of a page (eg. USB KVM scenarios), in
which case you may need to have a notification of when it changes (similar to,
eg., screen resolution).

https://codereview.chromium.org/20598008/diff/1/Source/modules/pointerevents/...
Source/modules/pointerevents/NavigatorPointerEvents.cpp:86: return 0;
On platforms where we don't know the answer it's better not to expose the API
then to return a potentially incorrect value.  This might mean we should have a
RuntimeEnabledFeature for this as for touch-action (which you'd need anyway if
you don't get simultaneous approval to implement and ship).

[sblom, 2013-10-10 20:54:17]

Updated this with the architecture that you helped me settle on.

[Rick Byers, 2013-10-10 22:12:17]

Looking good, almost there!

Please add a CL description including:
 - a link to the intent-to-implement-and-ship thread on blink-dev
 - a link to the spec where this API is defined

Also you should add a simple LayoutTest that just stores a value into
internals.settings and verifies it can read it out through
navigator.maxTouchPoints.

I think there may also be a test somewhere that needs to be updated with the new
public API surface area (although maybe we don't test that for Navigator).  You
should try running all the layout tests on your machine - it's surprisingly
quick (may need to exclude pixel tests - they can be tricky to get to match
expectations exactly).

https://codereview.chromium.org/20598008/diff/5001/Source/core/page/RuntimeEn...
File Source/core/page/RuntimeEnabledFeatures.in (right):

https://codereview.chromium.org/20598008/diff/5001/Source/core/page/RuntimeEn...
Source/core/page/RuntimeEnabledFeatures.in:73: ObjectFitPosition
status=experimental
I assume you didn't intend to change this, right?

https://codereview.chromium.org/20598008/diff/5001/Source/core/page/RuntimeEn...
Source/core/page/RuntimeEnabledFeatures.in:80: PointerEventsMaxTouchPoints
status=experimental
Please add a comment describing how this is platform specific for now, and
ideally linking to a bug that tracks removing the RuntimeEnabledFeature once
it's supported on all platforms (see the comment I recently added to this file
and Settings.in).

https://codereview.chromium.org/20598008/diff/5001/Source/core/page/Settings.in
File Source/core/page/Settings.in (right):

https://codereview.chromium.org/20598008/diff/5001/Source/core/page/Settings....
Source/core/page/Settings.in:127: maxTouchPoints type=int, initial=0
Please add a comment (see the other new comments in this file).  Eg. perhaps a
link to the pointer events spec where it defines the semantics of this value.

[sblom, 2013-10-17 17:44:27]

I'll have what I think is a final draft of both this patch and its chromium
companion this afternoon.