Remove Default Touch Radii in EventGenerator

Remove Default Touch Radii in EventGenerator

EventGenerator creates touch events in two different manners. Those that use TestTouchEvent set touch radii without exposing this to unit test classes. Chrome's gesture recognizer uses these radii when determining touch rects. This can lead to different views being targetting within unittests.

Remove the setting of these radii. Tests that wish to use larger can still create TouchEvents and call EventGenerator::Dispatch.

Committed: https://crrev.com/644b0306f7f23e5760728c0dc0d68923a8e13526
Cr-Commit-Position: refs/heads/master@{#292933}

[jonross, 2014-08-28 22:57:27]

jonross@chromium.org changed reviewers:
+ tdresser@chromium.org

[jonross, 2014-08-28 22:57:27]

Hi Tim,

As we discussed offline the method used for generating touch events was leading
to different gestures. Could you give the review a first look?

I ran tests that reference the methods offline. However I am not a committer, so
I cannot trigger try-bots. Could you trigger some?

[jonross, 2014-08-29 17:26:17]

jonross@chromium.org changed reviewers:
+ sadrul@chromium.org

[jonross, 2014-08-29 17:26:18]

Hi Sadrul,

Could you provide an owner review for this change?

While writing a unit test I noticed a difference in touch targets. Currently the
TestTouchEvent class is setting a radius on touches, triggering touch-fuzzing
code within the gesture recognizer. This was leading to me having different
results based on which EventGenerator methods I used.

I've ran the affected tests locally, and Tim has triggered trybots. Only
failures seen are unrelated in ASAN bot that has affected other changes.

Thanks,
Jon

[tdresser, 2014-08-29 17:32:29]

On 2014/08/29 17:26:18, jonross wrote:
> Hi Sadrul,
> 
> Could you provide an owner review for this change?
> 
> While writing a unit test I noticed a difference in touch targets. Currently
the
> TestTouchEvent class is setting a radius on touches, triggering touch-fuzzing
> code within the gesture recognizer. This was leading to me having different
> results based on which EventGenerator methods I used.
> 
> I've ran the affected tests locally, and Tim has triggered trybots. Only
> failures seen are unrelated in ASAN bot that has affected other changes.
> 
> Thanks,
> Jon

I'm not sure which makes more sense, getting rid of the radius here, or adding
it everywhere else.

I think this is reasonable though. LGTM.

[sadrul, 2014-09-02 14:29:31]

lgtm

[jonross, 2014-09-02 14:31:48]

The CQ bit was checked by jonross@chromium.org

[commit-bot: I haz the power, 2014-09-02 14:32:08]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/jonross@chromium.org/522453002/1

[commit-bot: I haz the power, 2014-09-02 15:42:49]

Committed patchset #1 (id:1) as ecdef09aebbe1b14de34662504e3fb569bc25756

[commit-bot: I haz the power, 2014-09-10 03:18:43]

Patchset 1 (id:??) landed as
https://crrev.com/644b0306f7f23e5760728c0dc0d68923a8e13526
Cr-Commit-Position: refs/heads/master@{#292933}