Always forward touch events to the TouchEventQueue

Always forward touch events to the TouchEventQueue

There's currently a shortcut that the RWHView can take with touch events
when there is no registered touch handler. This avoids the overhead
of passing the touch to the InputRouter/TouchEventQueue, but the
predicate for conditionally forwarding the events is racy. This can lead
to scenarios where a touch event stream is interrupted and the
pipeline only sees a partial touch sequence.

As the cost of always forwarding the events is relatively small, remove
this optimization entirely, instead relying on the TouchEventQueue to
filter events as appropriate.

BUG=461583
Committed: https://crrev.com/dfb809f04aa38b0fc0a6ce8c69649377b83a20df
Cr-Commit-Position: refs/heads/master@{#318243}

[jdduke (slow), 2015-02-25 19:05:37]

jdduke@chromium.org changed reviewers:
+ sadrul@chromium.org, tdresser@chromium.org

[jdduke (slow), 2015-02-25 19:05:37]

PTAL, thanks.

tdresser@: content/browser/renderer_host/input, ui/events
sadrul@: content/browser/renderer_host

[tdresser, 2015-02-25 19:17:56]

On 2015/02/25 19:05:37, jdduke wrote:
> PTAL, thanks.
> 
> tdresser@: content/browser/renderer_host/input, ui/events
> sadrul@: content/browser/renderer_host

content/browser/renderer_host/input, ui/events LGTM.

I'm a little worried that we're losing some valuable test coverage by removing
all calls to |ShouldForwardTouchEvent()|.
It probably wouldn't be worth adding a similar method for testing only, but I'm
not certain.

[jdduke (slow), 2015-02-25 22:51:07]

On 2015/02/25 19:17:56, tdresser wrote:
> On 2015/02/25 19:05:37, jdduke wrote:
> > PTAL, thanks.
> > 
> > tdresser@: content/browser/renderer_host/input, ui/events
> > sadrul@: content/browser/renderer_host
> 
> content/browser/renderer_host/input, ui/events LGTM.
> 
> I'm a little worried that we're losing some valuable test coverage by removing
> all calls to |ShouldForwardTouchEvent()|.
> It probably wouldn't be worth adding a similar method for testing only, but
I'm
> not certain.

I sprinkled in a few checks in the RWHVAura test to make sure touches are
getting forwarded (or not forwarded) to the renderer.

[tdresser, 2015-02-26 13:40:16]

On 2015/02/25 22:51:07, jdduke wrote:
> On 2015/02/25 19:17:56, tdresser wrote:
> > On 2015/02/25 19:05:37, jdduke wrote:
> > > PTAL, thanks.
> > > 
> > > tdresser@: content/browser/renderer_host/input, ui/events
> > > sadrul@: content/browser/renderer_host
> > 
> > content/browser/renderer_host/input, ui/events LGTM.
> > 
> > I'm a little worried that we're losing some valuable test coverage by
removing
> > all calls to |ShouldForwardTouchEvent()|.
> > It probably wouldn't be worth adding a similar method for testing only, but
> I'm
> > not certain.
> 
> I sprinkled in a few checks in the RWHVAura test to make sure touches are
> getting forwarded (or not forwarded) to the renderer.

Awesome, thanks.

[sadrul, 2015-02-26 14:23:55]

rs lgtm

[jdduke (slow), 2015-02-26 16:05:14]

The CQ bit was checked by jdduke@chromium.org

[jdduke (slow), 2015-02-26 16:05:15]

The patchset sent to the CQ was uploaded after l-g-t-m from
tdresser@chromium.org
Link to the patchset: https://codereview.chromium.org/954973003/#ps40001 (title:
"More validation")

[commit-bot: I haz the power, 2015-02-26 16:06:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/954973003/40001

[commit-bot: I haz the power, 2015-02-26 16:16:26]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-02-26 16:17:33]

Patchset 3 (id:??) landed as
https://crrev.com/dfb809f04aa38b0fc0a6ce8c69649377b83a20df
Cr-Commit-Position: refs/heads/master@{#318243}