Find the current fullscreen video element rather than the one on top of stack

Find the current fullscreen video element rather than the one on top of stack

When requestFullscren() is called, we don't know when willEnterFullscreen() will get called.

For android webview, requestFullscreen() might even result in no calls
to willenterfullscreen() if webapp doesn't support fullscreen video.
So before willEnterfullscreen is called, we should not use the provisional
video element on top of the stack to modify the layer tree.

Instead, we should wait for the provisional element to become current.
Therefore, we should call currentFullScreenElementFrom() on the fullscreen document.

BUG=389496
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=178714

[qinmin, 2014-07-16 18:54:26]

PTAL

[esprehn, 2014-07-16 19:29:09]

This doesn't make sense to me, how can the element end up full screen even if it
fails?

[qinmin, 2014-07-16 19:46:26]

On 2014/07/16 19:29:09, esprehn wrote:
> This doesn't make sense to me, how can the element end up full screen even if
it
> fails?

When requestFullscreen is called(), the element is added to the
FullscreenElementStack. 
But before willEnterFullscreen()/DidEnterFullscreen() is received, we should not
change the rendering model and modify layer transparency. However, the current
behavior on android violates this, since it assumes that browser process will
always send will/did-EnterFullscreen().
For webview, it is possible that web apps doesn't want to support fullscreen for
video. As a result, will/Did-Enterfullscreen() might not get called.
If we change the layer transparency and attach the video layer to the top of the
layer tree, then this will cause the issue.

[trchen, 2014-07-17 00:34:15]

On 2014/07/16 19:29:09, esprehn wrote:
> This doesn't make sense to me, how can the element end up full screen even if
it
> fails?

To be more accurate, the requestFullscreen() call never fail, but it can take
indefinite time for the browser to approve the request. In the meantime, the
provisional fullscreen element is pushed onto the stack, so we keep track and
activate fullscreen for it when we receive the approval. In the case that the
browser rejects fullscreen, we won't get any feedback, just as if the user never
clicked the okay button, and the provisional fullscreen element stay there
forever. The bug is that we took the provisional fullscreen element as current
fullscreen element without waiting for approval first. The CL is correcting
that.

[esprehn, 2014-07-17 00:48:54]

This needs a test then, and also please wrap your change description somewhere
sensible like 80-100 chars.

[qinmin, 2014-07-17 02:34:39]

On 2014/07/17 00:48:54, esprehn wrote:
> This needs a test then, and also please wrap your change description somewhere
> sensible like 80-100 chars.

Done. CL description changed, and added a layout test.
The layout test is not perfect, because we cannot simulate the situation that
the browser process doesn't send the will/did-enterFullscreen().
Instead, the layout test tests the case that requestFullscreen() fails due to
missing user gesture.

[qinmin, 2014-07-18 21:27:01]

On 2014/07/17 02:34:39, qinmin wrote:
> On 2014/07/17 00:48:54, esprehn wrote:
> > This needs a test then, and also please wrap your change description
somewhere
> > sensible like 80-100 chars.
> 
> Done. CL description changed, and added a layout test.
> The layout test is not perfect, because we cannot simulate the situation that
> the browser process doesn't send the will/did-enterFullscreen().
> Instead, the layout test tests the case that requestFullscreen() fails due to
> missing user gesture.

ping. esprehn@, is this good to go? thanks

[esprehn, 2014-07-21 21:21:57]

lgtm, it's a shame we have all this hackery.

https://codereview.chromium.org/399703002/diff/40001/LayoutTests/fullscreen/v...
File LayoutTests/fullscreen/video-fail-to-enter-full-screen.html (right):

https://codereview.chromium.org/399703002/diff/40001/LayoutTests/fullscreen/v...
LayoutTests/fullscreen/video-fail-to-enter-full-screen.html:2: <body>
Leave out body

[qinmin, 2014-07-22 01:55:01]

https://codereview.chromium.org/399703002/diff/40001/LayoutTests/fullscreen/v...
File LayoutTests/fullscreen/video-fail-to-enter-full-screen.html (right):

https://codereview.chromium.org/399703002/diff/40001/LayoutTests/fullscreen/v...
LayoutTests/fullscreen/video-fail-to-enter-full-screen.html:2: <body>
On 2014/07/21 21:21:57, esprehn wrote:
> Leave out body

Done.

[qinmin, 2014-07-22 01:55:04]

The CQ bit was checked by qinmin@chromium.org

[commit-bot: I haz the power, 2014-07-22 01:55:21]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/qinmin@chromium.org/399703002/60001

[commit-bot: I haz the power, 2014-07-22 03:01:24]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  linux_blink_dbg on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_dbg/builds/1...)
  linux_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_rel/builds/1...)
  mac_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/mac_blink_rel/builds/16446)
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/18092)
  mac_gpu_triggered_tests on tryserver.chromium.gpu
(http://build.chromium.org/p/tryserver.chromium.gpu/builders/mac_gpu_triggered...)
  win_gpu_triggered_tests on tryserver.chromium.gpu
(http://build.chromium.org/p/tryserver.chromium.gpu/builders/win_gpu_triggered...)

[commit-bot: I haz the power, 2014-07-22 03:27:01]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-07-22 03:27:02]

Try jobs failed on following builders:
  linux_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_rel/builds/1...)

[esprehn, 2014-07-22 03:36:47]

You're failing real tests:

virtual/android/fullscreen/video-fail-to-enter-full-screen.html
fullscreen/video-fail-to-enter-full-screen.html

[qinmin, 2014-07-22 16:46:58]

On 2014/07/22 03:36:47, esprehn wrote:
> You're failing real tests:
> 
> virtual/android/fullscreen/video-fail-to-enter-full-screen.html
> fullscreen/video-fail-to-enter-full-screen.html

The test expectations should be "RequestFullScreen" instead of
"RequestFullscreen". Fixed the mistake here.

[qinmin, 2014-07-22 21:44:23]

The CQ bit was checked by qinmin@chromium.org

[commit-bot: I haz the power, 2014-07-22 21:45:05]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/qinmin@chromium.org/399703002/80001

[commit-bot: I haz the power, 2014-07-23 00:27:01]

Change committed as 178714