Add fling velocity updates from gesture curves.

Add fling velocity updates from gesture curves.

As per crrev.com/13674008, WebGestureCurveTarget now supports fling velocity
notifications.  This patch hooks in the notifications from existing
WebGestureCurve implementations.

BUG=135975
NOTRY=true

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=197139

[jdduke (slow), 2013-04-25 19:01:39]

I've pulled out much of the gesture curve and fling velocity work from
crrev.com/13375020/.  Hopefully this makes review a bit more palatable... =/

[jamesr, 2013-04-25 20:19:41]

Sorry to be giving you the runaround.  I feel that this could probably be done
without needing to touch the WebKit APIs at all with a bit of refactoring, but I
realize that it's hard to know how to refactor something complex like this when
you're new the codebase and the refactoring wouldn't provide any immediate
benefit for the functionality you're trying to implement.

lgtm for this with a few nits.  It'll work and isn't horrible.

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/fling_animator...
File webkit/glue/fling_animator_impl_android.cc (right):

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/fling_animator...
webkit/glue/fling_animator_impl_android.cc:126: float scroll_to_velocity =
velocity / std::sqrt((float)diff_length_2);
nit: use c++ style casts, not c-style. so static_cast<float>(...)

although why don't you just call Length() on line 123 instead?

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/touch_fling_ge...
File webkit/glue/touch_fling_gesture_curve.cc (right):

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/touch_fling_ge...
webkit/glue/touch_fling_gesture_curve.cc:132: speed = 0;
nit: 0.f

[jdduke (slow), 2013-04-25 20:45:27]

No worries.  I would agree that there are likely (many) better ways to do this,
but as a newbie I'm still somewhat hesitant to push out non-trivial
refactorings. In any case, thanks.

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/fling_animator...
File webkit/glue/fling_animator_impl_android.cc (right):

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/fling_animator...
webkit/glue/fling_animator_impl_android.cc:126: float scroll_to_velocity =
velocity / std::sqrt((float)diff_length_2);
On 2013/04/25 20:19:41, jamesr wrote:
> nit: use c++ style casts, not c-style. so static_cast<float>(...)
> 
> although why don't you just call Length() on line 123 instead?
No reason other than that itch to optimize out sqrt's =/ This is a relatively
sparse operation so there's really no sense prematurely optimizing. I'll change
accordingly.

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/touch_fling_ge...
File webkit/glue/touch_fling_gesture_curve.cc (right):

https://codereview.chromium.org/14128010/diff/2001/webkit/glue/touch_fling_ge...
webkit/glue/touch_fling_gesture_curve.cc:132: speed = 0;
On 2013/04/25 20:19:41, jamesr wrote:
> nit: 0.f

Done.

[jdduke (slow), 2013-04-25 20:56:37]

Hmm, if you do know of a better way to route these notifications without going
through WebKit, I'm all ears.  This is the least pleasant part of the whole
overscroll effort; avoiding WebKit changes altogether would be ideal.

If you want, I can hold off commits for this and crrev.com/13674008/ while we
map out a different approach.

[jamesr, 2013-04-26 05:54:08]

> Hmm, if you do know of a better way to route these notifications without going
> through WebKit, I'm all ears.  This is the least pleasant part of the whole
> overscroll effort; avoiding WebKit changes altogether would be ideal.
> 
> If you want, I can hold off commits for this and crrev.com/13674008/ while we
> map out a different approach.

I started sketching out the design improvements I had in mind and they're more
involved than I initially thought (or I've just confused myself).  Go ahead with
this patch, if I find time to refactor I can remove the blink stuff and make it
easier to modify this in the future.

[commit-bot: I haz the power, 2013-04-29 13:56:33]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/14128010/17001

[commit-bot: I haz the power, 2013-04-29 18:33:32]

Retried try job too often on win_rel for step(s) browser_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2013-04-29 21:06:51]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/14128010/38001

[commit-bot: I haz the power, 2013-04-29 21:10:21]

Change committed as 197139