Dispatch a synthetic mouse exit to the widget handling mouse events when a widget grabs capture

Dispatch a synthetic ui::ET_MOUSE_EXITED event to the widget handling mouse
events when a different widget grabs capture.

This CL fixes the following scenario:
- An event handler sets some state as a result of the widget getting hovered.
- A different widget than the one handling mouse events grabs capture. (e.g.
  the menu)
Sending a ui::ET_MOUSE_EXITED event is necessary so that any state set as a 
result of the widget getting hovered is reset. In particular,
ui::ET_MOUSE_CAPTURE_CHANGED is not dispatched in this scenario.

BUG=382421
TEST=WidgetCaptureTest.MouseExitOnCaptureGrab

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=280072

[sky, 2014-06-23 20:39:52]

Sadrul is a better reviewer than I for this.

[sadrul, 2014-06-24 18:33:09]

There are some subtle changes on when exit/move events are sent in this CL. I
have sent out some tryjobs. I will do a review if this isn't breaking any
existing tests.

https://codereview.chromium.org/344793013/diff/80001/ui/aura/window_event_dis...
File ui/aura/window_event_dispatcher.cc (right):

https://codereview.chromium.org/344793013/diff/80001/ui/aura/window_event_dis...
ui/aura/window_event_dispatcher.cc:382: mouse_moved_handler_ != new_capture) {
Should you check for a non-null |new_capture| here?

[pkotwicz, 2014-06-24 21:18:43]

The CQ bit was checked by pkotwicz@chromium.org

[pkotwicz, 2014-06-24 21:18:44]

The CQ bit was unchecked by pkotwicz@chromium.org

[pkotwicz, 2014-06-25 15:16:35]

Sadrul, can you please take a look?

- I have fixed WidgetCaptureTest.MouseExitOnCaptureGrab so that it passes when
running in Debug too
- I have investigated why TouchEditableImplAuraTest.TouchCursorInTextfieldTest
fails on CrOS. The failure is because  WMTestHelper does not set an
aura::client::CursorClient, so
CompoundEventFilter::SetMouseEventsEnableStateOnEvent() never disables mouse
events.

[sadrul, 2014-06-26 12:59:35]

LGTM

https://codereview.chromium.org/344793013/diff/180001/ui/aura/window_event_di...
File ui/aura/window_event_dispatcher.cc (right):

https://codereview.chromium.org/344793013/diff/180001/ui/aura/window_event_di...
ui/aura/window_event_dispatcher.cc:205: mouse_pressed_handler_ = NULL;
Is it necessary to make the same changes in here?

In fact, is it possible to remove OnHostLostMouseGrab()? Looks like it's called
only from DesktopWindowTreeHostX11::DispatchEvent(), and it may make sense to
just call into OnOtherRootGotCapture() instead from there.

https://codereview.chromium.org/344793013/diff/180001/ui/views/widget/widget_...
File ui/views/widget/widget_interactive_uitest.cc (right):

https://codereview.chromium.org/344793013/diff/180001/ui/views/widget/widget_...
ui/views/widget/widget_interactive_uitest.cc:897: // TODO(pkotwicz): Make test
pass on CrOS.
I guess you need to create multiple WindowTreeHosts for the change in this CL

[pkotwicz, 2014-06-26 14:45:11]

The CQ bit was checked by pkotwicz@chromium.org

[pkotwicz, 2014-06-26 14:45:14]

The CQ bit was unchecked by pkotwicz@chromium.org

[pkotwicz, 2014-06-26 14:45:18]

https://codereview.chromium.org/344793013/diff/180001/ui/aura/window_event_di...
File ui/aura/window_event_dispatcher.cc (right):

https://codereview.chromium.org/344793013/diff/180001/ui/aura/window_event_di...
ui/aura/window_event_dispatcher.cc:205: mouse_pressed_handler_ = NULL;
I need to investigate why OnHostLostMouseGrab() is its own function. I have a
slight suspicion that it has to do with the bad things we do with capture during
tab dragging

[pkotwicz, 2014-06-26 14:45:20]

The CQ bit was checked by pkotwicz@chromium.org

[commit-bot: I haz the power, 2014-06-26 14:47:15]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/pkotwicz@chromium.org/344793013/180001

[commit-bot: I haz the power, 2014-06-26 18:00:09]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  win_chromium_rel on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/win_chromium_rel/buil...)

[commit-bot: I haz the power, 2014-06-26 19:30:57]

Change committed as 280072