Modify ui_controls::SendKeyPressNotifyWhenDone to use [NSApp postEvent: atStart:].

Modify ui_controls::SendKeyPressNotifyWhenDone to use [NSApp postEvent: atStart:].

Currently, in ui_controls_mac.mm, [NSApp postEvent: atStart:] is used to
dispatch events, except in SendKeyPressNotifyWhenDone, where [NSApp sendEvent:]
is used. The reason to use [NSApp sendEvent:] was the failure of
BrowserKeyEventsTest.CommandKeyEvents which has since been disabled
(crbug.com/81451). This CL modifies SendKeyPressNotifyWhenDone to use 
[NSApp postEvent: atStart:] instead. This is required since events sent using 
[NSApp sendEvent:] are not added to the event queue, hence the correct event is 
not reported by [NSApp currentEvent].

BUG=607429, 617110

[karandeepb, 2016-06-06 08:28:09]

Description was changed from

==========
ui_controls_mac.
==========

to

==========
Modify ui_controls::SendKeyPressNotifyWhenDone to use [NSApp postEvent:
atStart:].

Currently, in ui_controls_mac.mm, [NSApp postEvent: atStart:] is used to
dispatch events, except in SendKeyPressNotifyWhenDone, where [NSApp sendEvent:]
is used. The reason to use [NSApp sendEvent:] was the failure of
BrowserKeyEventsTest.CommandKeyEvents which has since been disabled
(crbug.com/81451). This CL modifies SendKeyPressNotifyWhenDone to use 
[NSApp postEvent: atStart:] instead. This is required since events sent using 
[NSApp sendEvent:] are not added to the event queue, hence the correct event is 
not reported by [NSApp currentEvent].

BUG=607429, 617110
==========

[karandeepb, 2016-06-06 09:46:24]

karandeepb@chromium.org changed reviewers:
+ thakis@chromium.org

[karandeepb, 2016-06-06 09:46:24]

PTAL @thakis. Require this for crrev.com/2033433006.

[karandeepb, 2016-06-07 05:50:49]

ping @thakis. Thanks.

[karandeepb, 2016-06-07 07:31:19]

Patchset #2 (id:20001) has been deleted

[Nico, 2016-06-09 15:13:33]

Wait, the motivation for removing a comment saying "We can't just $foo, it makes
a test fail" is "the test is already disabled anyhow"?

This code hasn't been touched in 6 years and has explicit comments saying "don't
do foo" and this removes the comments and changes the code to do foo. That needs
a better justification for why this is a safe change imho.

[karandeepb, 2016-06-10 08:15:22]

On 2016/06/09 15:13:33, Nico wrote:
> Wait, the motivation for removing a comment saying "We can't just $foo, it
makes
> a test fail" is "the test is already disabled anyhow"?
> 
> This code hasn't been touched in 6 years and has explicit comments saying
"don't
> do foo" and this removes the comments and changes the code to do foo. That
needs
> a better justification for why this is a safe change imho.

[karandeepb, 2016-06-10 08:15:50]

On 2016/06/10 08:15:22, karandeepb wrote:
> On 2016/06/09 15:13:33, Nico wrote:
> > Wait, the motivation for removing a comment saying "We can't just $foo, it
> makes
> > a test fail" is "the test is already disabled anyhow"?
> > 
> > This code hasn't been touched in 6 years and has explicit comments saying
> "don't
> > do foo" and this removes the comments and changes the code to do foo. That
> needs
> > a better justification for why this is a safe change imho.

The way I read it was using [NSApp postEvent: atStart:] was the correct thing to
do, but due to crbug.com/49270, we had to use [NSApp sendEvent:]. Since the test
was disabled, I thought it should be ok to use [NSApp postEvent: atStart:] now. 

Anyways, I'll look into what causes the test failure with [NSApp postEvent:
atStart:] (crbug.com/49270) and follow up.

[karandeepb, 2016-06-29 00:52:37]

So I did some investigation but couldn't find the root cause behind why
postEvent doesn't work. Have commented on the bug with my findings -
https://bugs.chromium.org/p/chromium/issues/detail?id=49270#c9