Remove TouchAdjustment.

Remove TouchAdjustment.

TouchAdjustment adjusts the hit test rects to nearby elements so
that when touch targets are too small they can still be hit by users. Instead
applications should just make their touch targets big enough, and if they
need adjustment it should be implemented at the framework level.

R=ojan@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/21ce97a2c082187f587e1ca8c0d24c4472b81637

[esprehn, 2014-12-03 02:57:23]

esprehn@chromium.org changed reviewers:
+ ojan@chromium.org, rafaelw@chromium.org

[ojan, 2014-12-03 03:04:23]

+rbyers to let us know if removing this is crazy or a good idea.

otherwise, lgtm

[Rick Byers, 2014-12-03 14:36:57]

On 2014/12/03 03:04:23, ojan wrote:
> +rbyers to let us know if removing this is crazy or a good idea.
> 
> otherwise, lgtm

Hah, I wrote a big response assuming this discussion was about blink, but of
course it's about sky ;-).  

TL;dr - yes, it probably has no place in the core of sky, it's (probably) a hack
to work around UI that isn't designed to be touch friendly.  We're going to try
an experiment to disable it on mobile websites here http://crbug.com/417534. 
I'd also like to expose a rect-based hit-test API to the web (like IE has:
http://ie.microsoft.com/TESTdrive/HTML5/HitTest/Default.html) so frameworks
could reasonably provide this functionality themselves if they really want it.

However, I might be wrong about this.  Does iOS or Android have any built-in
notion (outside of webview of course) of fuzzy hit-testing?  Apple has
traditionally been unwilling to discuss this publicly due to IP issues.

[esprehn, 2014-12-03 18:51:31]

Committed patchset #1 (id:1) manually as
21ce97a2c082187f587e1ca8c0d24c4472b81637 (presubmit successful).

[ojan, 2014-12-03 21:20:45]

On 2014/12/03 at 14:36:57, rbyers wrote:
> On 2014/12/03 03:04:23, ojan wrote:
> > +rbyers to let us know if removing this is crazy or a good idea.
> > 
> > otherwise, lgtm
> 
> Hah, I wrote a big response assuming this discussion was about blink, but of
course it's about sky ;-).  

Heh. Sorry.

> TL;dr - yes, it probably has no place in the core of sky, it's (probably) a
hack to work around UI that isn't designed to be touch friendly.  We're going to
try an experiment to disable it on mobile websites here http://crbug.com/417534.
 I'd also like to expose a rect-based hit-test API to the web (like IE has:
http://ie.microsoft.com/TESTdrive/HTML5/HitTest/Default.html) so frameworks
could reasonably provide this functionality themselves if they really want it.
>

Yeah, that's what we were thinking, but I know you've thought about this stuff a
lot more than we have.

> However, I might be wrong about this.  Does iOS or Android have any built-in
notion (outside of webview of course) of fuzzy hit-testing?  Apple has
traditionally been unwilling to discuss this publicly due to IP issues.

Not sure.

[esprehn, 2014-12-04 07:26:38]

On 2014/12/03 at 21:20:45, ojan wrote:
> On 2014/12/03 at 14:36:57, rbyers wrote:
> > ...
> 
> > However, I might be wrong about this.  Does iOS or Android have any built-in
notion (outside of webview of course) of fuzzy hit-testing?  Apple has
traditionally been unwilling to discuss this publicly due to IP issues.
> 
> Not sure.

Android lets you implement this yourself:
https://github.com/android/platform_frameworks_base/blob/master/core/java/and...

The framework also supports some slop features for gestures like dragging from
the edge. See all the slop properties in ViewConfiguration:
http://developer.android.com/reference/android/view/ViewConfiguration.html#ge...

iOS also lets you do this yourself:
http://stackoverflow.com/questions/808503/uibutton-making-the-hit-area-larger...

and some of the built in widgets do it too:
http://stackoverflow.com/questions/9079907/why-does-uinavigationbar-steal-tou...

but neither platform has a general concept of touch adjustment.