Make ui::ThreadedTransformTransition::OnAbort correctly check whether the animation has started

Make ui::ThreadedTransformTransition::OnAbort correctly check whether the animation has started

This was incorrectly treating a non-zero animation id as an
indication that the animation had started.

BUG=258037
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=210753

[ajuma, 2013-07-09 18:58:39]

vollick@, PTAL.

This fixes the janky screen-lock cancellation animation described in
http://crbug.com/258037. 

Note that antrim@ investigated this and had a fix
(https://chromiumcodereview.appspot.com/18801008) that made threaded animations
never update their delegate when aborted; this effectively relied on the fact
that cc's animation framework periodically updates the main-thread values of
animated properties. However, when testing that approach on-device (on a Pixel),
the animation was still noticeably janky (though less janky than before); the
problem is that the main-thread values can significantly lag behind the
impl-thread values (since the main-thread values are only updated once per
commit, so if we haven't committed in a while these values are stale).
Consequently, when we abort a threaded animation that has started, we do in fact
need to update the delegate.

[Ian Vollick, 2013-07-10 01:37:49]

On 2013/07/09 18:58:39, ajuma wrote:
> vollick@, PTAL.
> 
> This fixes the janky screen-lock cancellation animation described in
> http://crbug.com/258037. 
> 
> Note that antrim@ investigated this and had a fix
> (https://chromiumcodereview.appspot.com/18801008) that made threaded
animations
> never update their delegate when aborted; this effectively relied on the fact
> that cc's animation framework periodically updates the main-thread values of
> animated properties. However, when testing that approach on-device (on a
Pixel),
> the animation was still noticeably janky (though less janky than before); the
> problem is that the main-thread values can significantly lag behind the
> impl-thread values (since the main-thread values are only updated once per
> commit, so if we haven't committed in a while these values are stale).
> Consequently, when we abort a threaded animation that has started, we do in
fact
> need to update the delegate.

lgtm!

[commit-bot: I haz the power, 2013-07-10 02:12:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ajuma@chromium.org/18660005/1

[commit-bot: I haz the power, 2013-07-10 05:07:52]

Change committed as 210753