Enable setting capture to a non-toplevel widget when the widget is activated on Linux

Enable setting capture to a non-toplevel widget when the widget is activated on
Linux

This CL also makes DesktopWindowTreeHostX11 always synchronously notify the
aura::Window which lost capture when the capture release is requested. (as
opposed to when capture is actually lost) Previously,
CaptureClient::SetCapture() would synchronously notify the aura::Window which
lost capture.
DesktopWindowTreeHostX11::Minimize() would asynchronousy notify the aura::Window
which lost capture.

BUG=319516, 392480
TEST=WidgetCaptureTest.SetCaptureToNonToplevel

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=283034

[pkotwicz, 2014-06-30 21:59:20]

Sadrul, can you please take a look?

Making capture "look" synchronous makes it easier to test the code.

Eventually, I want to make changing capture from one widget to another actually
synchronous. (one idea I have is a single X window which grabs capture whenever
a views::Widget requests capture)

[sadrul, 2014-07-02 15:20:28]

What broken feature in chrome is this CL fixing? Can you link to a bug?

[pkotwicz, 2014-07-14 14:57:53]

Sadrul, ping?
I have linked this CL with the two bugs that it is blocking.

[sadrul, 2014-07-14 16:58:01]

Looks good. A question:

https://codereview.chromium.org/334933008/diff/60001/ui/views/widget/desktop_...
File ui/views/widget/desktop_aura/desktop_window_tree_host_x11.cc (left):

https://codereview.chromium.org/334933008/diff/60001/ui/views/widget/desktop_...
ui/views/widget/desktop_aura/desktop_window_tree_host_x11.cc:825:
x11_capture_.reset();
Can you explain this change?

https://codereview.chromium.org/334933008/diff/60001/ui/views/widget/widget_i...
File ui/views/widget/widget_interactive_uitest.cc (right):

https://codereview.chromium.org/334933008/diff/60001/ui/views/widget/widget_i...
ui/views/widget/widget_interactive_uitest.cc:1016: virtual void
OnWidgetActivationChanged(Widget* widget, bool active) OVERRIDE {
// WidgetObserver:

[pkotwicz, 2014-07-14 18:25:20]

Sadrul, can you please take another look?

https://codereview.chromium.org/334933008/diff/60001/ui/views/widget/desktop_...
File ui/views/widget/desktop_aura/desktop_window_tree_host_x11.cc (left):

https://codereview.chromium.org/334933008/diff/60001/ui/views/widget/desktop_...
ui/views/widget/desktop_aura/desktop_window_tree_host_x11.cc:825:
x11_capture_.reset();
This line in particular made it pointless to grab capture when a non-toplevel
widget got activated (because the toplevel widget would lose focus killing the
capture which was just set)

Note that the order of events is:
1) old window loses activation
2) new window gains activation
3) old window loses focus
4) new window gains focus

I think that the intent of the code is to release capture when we are
deactivated. I have modified HandleNativeWidgetActivationChanged() to do just
that

The reason that I changed x11_capture_.reset() to ReleaseCapture() in this CL is
that the test that I wrote would pass with or without this CL (due to the fact
that the effect of x11_capture_.reset() is asynchronous. Using ReleaseCapture()
makes the test actually useful.

[sadrul, 2014-07-14 18:32:02]

LGTM

[pkotwicz, 2014-07-14 18:57:03]

The CQ bit was checked by pkotwicz@chromium.org

[commit-bot: I haz the power, 2014-07-14 18:57:30]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/pkotwicz@chromium.org/334933008/110001

[pkotwicz, 2014-07-14 18:59:15]

The CQ bit was unchecked by pkotwicz@chromium.org

[pkotwicz, 2014-07-14 19:15:34]

The CQ bit was checked by pkotwicz@chromium.org

[commit-bot: I haz the power, 2014-07-14 19:17:13]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/pkotwicz@chromium.org/334933008/130001

[commit-bot: I haz the power, 2014-07-14 21:47:50]

Change committed as 283034