Add EventFilter for LauncherTooltip.

Add EventFilter for LauncherTooltip.

Now it monitors mouse/touch/gesture events and hides the tooltip if necessary.

BUG=137678
TEST=manually verified && aura_shell_unittests passed


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=148079

[Daniel Erat, 2012-07-19 14:51:18]

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
File ash/launcher/launcher_tooltip_manager.cc (right):

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:251: // Because this mouse event may
arrive to |view_|, here we just schedules
nit: s/schedules/schedule/

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:252: // the closing event rather than
directly calls Close().
nit: s/calls/calling/

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:270: // Because this mouse event may
arrive to |view_|, here we just schedules
nit: same updates as described above

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:307:
Shell::GetInstance()->AddEnvEventFilter(this);
ShowInternal() looks like it gets called by ShowImmediately().  Is every call
guaranteed to be paired with exactly one call to Close()?

[Jun Mukai, 2012-07-20 05:00:11]

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
File ash/launcher/launcher_tooltip_manager.cc (right):

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:251: // Because this mouse event may
arrive to |view_|, here we just schedules
On 2012/07/19 14:51:18, Daniel Erat wrote:
> nit: s/schedules/schedule/

Done.

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:252: // the closing event rather than
directly calls Close().
On 2012/07/19 14:51:18, Daniel Erat wrote:
> nit: s/calls/calling/

Done.

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:270: // Because this mouse event may
arrive to |view_|, here we just schedules
On 2012/07/19 14:51:18, Daniel Erat wrote:
> nit: same updates as described above

Done.

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:307:
Shell::GetInstance()->AddEnvEventFilter(this);
On 2012/07/19 14:51:18, Daniel Erat wrote:
> ShowInternal() looks like it gets called by ShowImmediately().  Is every call
> guaranteed to be paired with exactly one call to Close()?

fixed the code a bit.  Close() is called in the destructor which removes env
event filter, so it would be fine.  Also added to handle in WindowClosing().

[Jun Mukai, 2012-07-20 05:06:46]

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
File ash/launcher/launcher_tooltip_manager.cc (right):

http://codereview.chromium.org/10810011/diff/1/ash/launcher/launcher_tooltip_...
ash/launcher/launcher_tooltip_manager.cc:307:
Shell::GetInstance()->AddEnvEventFilter(this);
On 2012/07/20 05:00:11, Jun Mukai wrote:
> On 2012/07/19 14:51:18, Daniel Erat wrote:
> > ShowInternal() looks like it gets called by ShowImmediately().  Is every
call
> > guaranteed to be paired with exactly one call to Close()?
> 
> fixed the code a bit.  Close() is called in the destructor which removes env
> event filter, so it would be fine.  Also added to handle in WindowClosing().

sorry, I changed my mind and uploaded another patchset.  It calls
AddEnvEventFilter in the constuctor and call Remove in the destructor, so it
would guarantee the behavior.  Then I added guards of if (widget_ &&
widget_->IsVisible())... in the event filter methods.

[Jun Mukai, 2012-07-23 14:23:24]

dan, can you take another look at this?

[Daniel Erat, 2012-07-23 15:45:53]

LGTM

Thanks, the observer code looks much better!

[commit-bot: I haz the power, 2012-07-24 09:08:57]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/mukai@chromium.org/10810011/7002

[commit-bot: I haz the power, 2012-07-24 11:15:14]

Change committed as 148079