Always notify the MediaPlayer of any seek.

Always notify the MediaPlayer of any seek.

This change removes short circuit logic for seeking to the current position.
There are certain situations, like seeking to a truncated duration, where the
HTMLMediaElement is not in the best position to determine whether a call to
the MediaPlayer is necessary or not. This change allows the
MediaPlayer/WebMediaPlayer implementation to determine what the best course
of action should be. This also makes the seeking behavior more spec
compliant because even a seek to the current position should have the
seeking attribute be true until a stable state occurs. This is guaranteed
when calling into the MediaPlayer and allows all seeks to be treated the
same at the HTMLMediaElement level.

BUG=266631
TEST=LayoutTests/media/video-seek-to-current-position.html

[acolwell GONE FROM CHROMIUM, 2014-06-17 21:52:36]

https://codereview.chromium.org/340553004/diff/1/LayoutTests/media/media-cont...
File LayoutTests/media/media-controller-effective-playback-rate.html (right):

https://codereview.chromium.org/340553004/diff/1/LayoutTests/media/media-cont...
LayoutTests/media/media-controller-effective-playback-rate.html:16:
waitForEventOnce("canplay",canplay);
This change and the other one like it is needed because this test assumes that
canplay will only fire once. Part of the media controller's "bringing the
element up to speed" algorithm issues a seek on the element in response to
play(). Now that the media engine is being notified of all seeks, a readyState
transition to HAVE_METADATA can occur which opens up the possibility for another
"canplay" event to fire.

This change simply changes to test to reflect the original intent which was "on
the first canplay event call canplay".

[adamk, 2014-06-17 23:08:40]

lgtm, but philipj definitely has a better understanding of this code at this
point

[philipj_slow, 2014-06-18 09:03:41]

LGTM with enthusiasm. This also fixes a test in web-platform-tests I wrote
targeting this bug:
https://github.com/w3c/web-platform-tests/blob/master/html/semantics/embedded...

https://codereview.chromium.org/340553004/diff/1/LayoutTests/media/video-seek...
File LayoutTests/media/video-seek-to-current-position.html (right):

https://codereview.chromium.org/340553004/diff/1/LayoutTests/media/video-seek...
LayoutTests/media/video-seek-to-current-position.html:38: if (seekedCount == 3)
{
I think this should say 2?

[acolwell GONE FROM CHROMIUM, 2014-06-18 17:02:41]

https://codereview.chromium.org/340553004/diff/1/LayoutTests/media/video-seek...
File LayoutTests/media/video-seek-to-current-position.html (right):

https://codereview.chromium.org/340553004/diff/1/LayoutTests/media/video-seek...
LayoutTests/media/video-seek-to-current-position.html:38: if (seekedCount == 3)
{
On 2014/06/18 09:03:41, philipj wrote:
> I think this should say 2?

Done.

[acolwell GONE FROM CHROMIUM, 2014-06-18 17:02:46]

The CQ bit was checked by acolwell@chromium.org

[commit-bot: I haz the power, 2014-06-18 17:03:08]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/acolwell@chromium.org/340553004/20001

[commit-bot: I haz the power, 2014-06-18 18:53:42]

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

[commit-bot: I haz the power, 2014-06-18 18:53:43]

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...)

[philipj_slow, 2014-07-17 11:49:35]

Is this still good for landing, or are the failing tests indicative of a real
problem?

[acolwell GONE FROM CHROMIUM, 2014-07-17 14:32:25]

On 2014/07/17 11:49:35, philipj wrote:
> Is this still good for landing, or are the failing tests indicative of a real
> problem?

There was a real problem, but I think it might be resolved by now. I just
haven't had a chance to rebase and try again. It is on my TODO list.

[Srirama, 2014-07-25 15:02:47]

On 2014/07/17 14:32:25, acolwell wrote:
> On 2014/07/17 11:49:35, philipj wrote:
> > Is this still good for landing, or are the failing tests indicative of a
real
> > problem?
> 
> There was a real problem, but I think it might be resolved by now. I just
> haven't had a chance to rebase and try again. It is on my TODO list.

Hi Aaron,
If you are busy with other priority work, can i try landing this one if you
don't mind?

[acolwell GONE FROM CHROMIUM, 2014-07-25 16:02:30]

On 2014/07/25 15:02:47, Srirama wrote:
> On 2014/07/17 14:32:25, acolwell wrote:
> > On 2014/07/17 11:49:35, philipj wrote:
> > > Is this still good for landing, or are the failing tests indicative of a
> real
> > > problem?
> > 
> > There was a real problem, but I think it might be resolved by now. I just
> > haven't had a chance to rebase and try again. It is on my TODO list.
> 
> Hi Aaron,
> If you are busy with other priority work, can i try landing this one if you
> don't mind?

Sure. I think this patch is almost there. I rebased yesterday and it appears
that there is still an issue with some media controller tests. It isn't clear to
me if there is a real problem or if the test just needs to be updated because it
was relying on something that it should have. Please upload my original patch
first and then upload any followup changes you make. That will make it easier to
see what you've changed from what I have here.

Thanks.