Make MediaSession a runtime-enabled feature on Desktop.

Make MediaSession a runtime-enabled feature on Desktop.

At the moment, it's only compiled on Android, this CL compiles it on all platforms
but it's only enabled by default on Android. Other platforms have to pass
--enable-default-media-session flag to the command line.

This CL move all the media session related files from content/browser/media/android/
to content/browser/media/session/ and tear down the Android specific code from
MediaSession into a delegate. The Desktop delegate simply pauses other tabs when a
media session gets activated in a tab.

It is also moving the MediaSessionController handling to a separate class so it
can move from MediaWebContentsObserverAndroid to MediaWebContentsObserver easily.

BUG=587471
Committed: https://crrev.com/91873409e26ae1547dcd759735348884823f66ad
Cr-Commit-Position: refs/heads/master@{#380850}

[mlamouri (slow - plz ping), 2016-02-17 15:43:41]

Description was changed from

==========
WIP: Enable MediaSession on all platforms.

... but keep it behind a flag for non-Android.

BUG=<TODO>
==========

to

==========
WIP: Enable MediaSession on all platforms.

... but keep it behind a flag for non-Android.

BUG=587471
==========

[mlamouri (slow - plz ping), 2016-02-17 19:27:37]

Description was changed from

==========
WIP: Enable MediaSession on all platforms.

... but keep it behind a flag for non-Android.

BUG=587471
==========

to

==========
Make MediaSession a runtime-enabled feature on Desktop.

At the moment, it's only compiled on Android, this CL compiles it on all
platform
but it's only enabled by default on Android. Other platforms have to pass
--enable-default-media-session flag to the command line.

This CL move all the media session related files from
content/browser/media/android/
to content/browser/media/session/ and tear down the Android specific code from
MediaSession into a delegate. The Desktop delegate simply pauses other tab when
a
media session gets activated in a tab.

It is also moving the MediaSessionController handling to a separate class so it
can move from MediaWebContentsObserverAndroid to MediaWebContentsObserver
easily.

BUG=587471
==========

[mlamouri (slow - plz ping), 2016-02-17 19:30:11]

mlamouri@chromium.org changed reviewers:
+ avayvod@chromium.org, dalecurtis@chromium.org, jochen@chromium.org

[mlamouri (slow - plz ping), 2016-02-17 19:30:12]

Could you PTAL at:

avayvod@:
 - content/browser/media/session/

dalecurtis@:
 - content/browser/media/
 - media/


jochen@:
 - content/browser/web_contents/
 - content/public/

Thanks \o/

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/s...
File content/browser/media/session/media_session_browsertest.cc (right):

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/s...
content/browser/media/session/media_session_browsertest.cc:1297: class
DefaultMediaSessionBrowserTest : public content::ContentBrowserTest {
I'm going to move this test to media_session_delegate_default_browsertest.cc but
I do not have the time today and the CL is otherwise ready to be looked at.

[DaleCurtis, 2016-02-17 20:16:54]

Also, note this won't work on desktop for things like WebAudio or Pepper, you
need to make the ducking happen at the AudioRendererHost/WebContents::SetMuted()
level instead if you want this to work there.

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
File content/browser/media/media_web_contents_observer.cc (right):

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.cc:41:
session_controllers_manager_.Clear(render_frame_host);
We are now mixing a pattern of a delegate, 2 managers, and a controller for
MediaSession. It has a bit of smell to it :( It seems these layers should be
folded away, not added too. I don't have a concrete suggestion though. My
thought at the start of this all was that MediaSessionManager should do more
than it does today.

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.cc:99: // |has_video| forced
to true since the value doesn't matter at present.
This no longer does quite the right thing, we don't want to force another
OnMediaPlaying() here since it will trip the power save blocker additions.
Previously it only ran the android variant of this code, now it runs both (power
save blocker + session controller addition).

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.cc:102: return
media_session_map_.find(MediaPlayerId(
No media_session_map_ anymore.

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
File content/browser/media/media_web_contents_observer.h (right):

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.h:44: #if defined(OS_ANDROID)
Can you keep this on MWCOA? I'd like to avoid reintroducing any #ifdefs to this
class. Perhaps by exposing a protected method that derived classes can use to
access the session map?

[mlamouri (slow - plz ping), 2016-02-23 20:27:23]

The CQ bit was checked by mlamouri@chromium.org to run a CQ dry run

[mlamouri (slow - plz ping), 2016-02-23 20:27:55]

Review comments applied. PTAL.

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
File content/browser/media/media_web_contents_observer.cc (right):

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.cc:41:
session_controllers_manager_.Clear(render_frame_host);
On 2016/02/17 at 20:16:54, DaleCurtis wrote:
> We are now mixing a pattern of a delegate, 2 managers, and a controller for
MediaSession. It has a bit of smell to it :( It seems these layers should be
folded away, not added too. I don't have a concrete suggestion though. My
thought at the start of this all was that MediaSessionManager should do more
than it does today.

The delegate is really because of platform dependant implementations. This thing
is the first "manager", though I guess you consider content::MediaSession as a
manager?

Otherwise, regarding MediaSessionControllersManager, I think it helps make the
code simpler in MediaWebContentsObserver. It also allow re-use in
MediaWebContentsObserverAndroid. This said, I would be willing to refactor that
if you have some ideas. Maybe in another CL though to not make this even larger.

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.cc:99: // |has_video| forced
to true since the value doesn't matter at present.
On 2016/02/17 at 20:16:54, DaleCurtis wrote:
> This no longer does quite the right thing, we don't want to force another
OnMediaPlaying() here since it will trip the power save blocker additions.
Previously it only ran the android variant of this code, now it runs both (power
save blocker + session controller addition).

This is actually neatly solved with the MediaSessionControllersManager: just
call .RequestPlay() instead of going back to ::OnMediayPlaying().

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.cc:102: return
media_session_map_.find(MediaPlayerId(
On 2016/02/17 at 20:16:54, DaleCurtis wrote:
> No media_session_map_ anymore.

Done.

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
File content/browser/media/media_web_contents_observer.h (right):

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/m...
content/browser/media/media_web_contents_observer.h:44: #if defined(OS_ANDROID)
On 2016/02/17 at 20:16:54, DaleCurtis wrote:
> Can you keep this on MWCOA? I'd like to avoid reintroducing any #ifdefs to
this class. Perhaps by exposing a protected method that derived classes can use
to access the session map?

Done.

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/s...
File content/browser/media/session/media_session_browsertest.cc (right):

https://codereview.chromium.org/1698933004/diff/60001/content/browser/media/s...
content/browser/media/session/media_session_browsertest.cc:1297: class
DefaultMediaSessionBrowserTest : public content::ContentBrowserTest {
On 2016/02/17 at 19:30:12, Mounir Lamouri wrote:
> I'm going to move this test to media_session_delegate_default_browsertest.cc
but I do not have the time today and the CL is otherwise ready to be looked at.

Done.

[mlamouri (slow - plz ping), 2016-02-23 20:28:42]

Oh, I forgot, regarding pepper and Web Audio, the intent is just to move the
code to Desktop and have it behind a flag. Having the feature fully work on
Desktop is for later ;)

[commit-bot: I haz the power, 2016-02-23 20:28:44]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1698933004/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1698933004/100001

[commit-bot: I haz the power, 2016-02-23 20:56:56]

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

[commit-bot: I haz the power, 2016-02-23 20:56:57]

Dry run: Try jobs failed on following builders:
  android_compile_dbg on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_comp...)
  cast_shell_linux on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)
  linux_chromium_chromeos_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  linux_chromium_gn_chromeos_rel on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[DaleCurtis, 2016-02-24 01:49:48]

Didn't have time to make it through, but overall looks good aside from the
matroska doll of patterns as discussed in my previous comment. I'm fine with
resolving that in a follow up CL.

Looks like you have some compile errors, so I'll take another look over the
details tomorrow.

[mlamouri (slow - plz ping), 2016-02-25 12:06:43]

The CQ bit was checked by mlamouri@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-02-25 12:07:00]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1698933004/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1698933004/160001

[mlamouri (slow - plz ping), 2016-02-25 12:07:11]

I'm taking a risk but I believe the bots should be green now :)

[commit-bot: I haz the power, 2016-02-25 12:49:30]

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

[commit-bot: I haz the power, 2016-02-25 12:49:31]

Dry run: Try jobs failed on following builders:
  android_clang_dbg_recipe on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_clan...)

[mlamouri (slow - plz ping), 2016-02-25 20:24:16]

FYI, this is passing try apart from Windows component builds. thakis@ was nice
enough to look into this but it's unfortunately not a trivial solution. It will
require some kind of hack (I'm willing to simply not CONTENT_EXPORT on Windows).
See https://crbug.com/589840

The CL is otherwise ready to review.

[DaleCurtis, 2016-02-26 03:22:50]

lgtm % a few nits, nice cleanup. I didn't review the actual MediaSession stuff
too closely, deferring to avayod@ for that.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/media_web_contents_observer.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/media_web_contents_observer.h:53:
MediaSessionControllersManager& session_controllers_manager() {
non-const& aren't allowed by the style guide.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/session/media_session.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session.h:113: void
SetDelegateForTest(scoped_ptr<MediaSessionDelegate> delegate);
You need to use ForTesting or ForTests otherwise the PRESUBMIT won't catch them.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/session/media_session_controllers_manager.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_controllers_manager.h:21: using
MediaPlayerId = std::pair<RenderFrameHost*, int>;
Why not include web_contents_observer.h to get this. Or put it somewhere else
that WCO can include?

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_controllers_manager.h:25: class
MediaSessionControllersManager {
You'll want CONTENT_EXPORT if you're planning to write a unittest for this.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_controllers_manager.h:48: static
bool IsDefaultMediaSessionEnabled();
Keep as a .cc local method?

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/session/media_session_delegate_android.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_delegate_android.h:17: class
MediaSessionDelegateAndroid : public MediaSessionDelegate {
Ditto on the export. Surprised your browser tests compile.

[jochen (gone - plz use gerrit), 2016-02-26 12:18:22]

lgtm

[mlamouri (slow - plz ping), 2016-02-26 12:25:37]

Description was changed from

==========
Make MediaSession a runtime-enabled feature on Desktop.

At the moment, it's only compiled on Android, this CL compiles it on all
platform
but it's only enabled by default on Android. Other platforms have to pass
--enable-default-media-session flag to the command line.

This CL move all the media session related files from
content/browser/media/android/
to content/browser/media/session/ and tear down the Android specific code from
MediaSession into a delegate. The Desktop delegate simply pauses other tab when
a
media session gets activated in a tab.

It is also moving the MediaSessionController handling to a separate class so it
can move from MediaWebContentsObserverAndroid to MediaWebContentsObserver
easily.

BUG=587471
==========

to

==========
Make MediaSession a runtime-enabled feature on Desktop.

At the moment, it's only compiled on Android, this CL compiles it on all
platforms
but it's only enabled by default on Android. Other platforms have to pass
--enable-default-media-session flag to the command line.

This CL move all the media session related files from
content/browser/media/android/
to content/browser/media/session/ and tear down the Android specific code from
MediaSession into a delegate. The Desktop delegate simply pauses other tabs when
a
media session gets activated in a tab.

It is also moving the MediaSessionController handling to a separate class so it
can move from MediaWebContentsObserverAndroid to MediaWebContentsObserver
easily.

BUG=587471
==========

[mlamouri (slow - plz ping), 2016-02-26 12:25:47]

The CQ bit was checked by mlamouri@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-02-26 12:25:51]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1698933004/220001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1698933004/220001

[mlamouri (slow - plz ping), 2016-02-26 12:26:18]

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/media_web_contents_observer.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/media_web_contents_observer.h:53:
MediaSessionControllersManager& session_controllers_manager() {
On 2016/02/26 at 03:22:50, DaleCurtis wrote:
> non-const& aren't allowed by the style guide.

Done.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/session/media_session.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session.h:113: void
SetDelegateForTest(scoped_ptr<MediaSessionDelegate> delegate);
On 2016/02/26 at 03:22:50, DaleCurtis wrote:
> You need to use ForTesting or ForTests otherwise the PRESUBMIT won't catch
them.

Done.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/session/media_session_controllers_manager.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_controllers_manager.h:21: using
MediaPlayerId = std::pair<RenderFrameHost*, int>;
On 2016/02/26 at 03:22:50, DaleCurtis wrote:
> Why not include web_contents_observer.h to get this. Or put it somewhere else
that WCO can include?

I don't really like the idea of including a file like that for a type def but I
guess it's probably cleaner. I've added a comment saying that it's needed for
MediaPlayerId.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_controllers_manager.h:25: class
MediaSessionControllersManager {
On 2016/02/26 at 03:22:50, DaleCurtis wrote:
> You'll want CONTENT_EXPORT if you're planning to write a unittest for this.

There are no tests for this. I don't think adding yet more stuff to this CL will
do any good. Opened a bug to add unit tests for M51: https://crbug.com/590183

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_controllers_manager.h:48: static
bool IsDefaultMediaSessionEnabled();
On 2016/02/26 at 03:22:50, DaleCurtis wrote:
> Keep as a .cc local method?

Done.

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
File content/browser/media/session/media_session_delegate_android.h (right):

https://codereview.chromium.org/1698933004/diff/200001/content/browser/media/...
content/browser/media/session/media_session_delegate_android.h:17: class
MediaSessionDelegateAndroid : public MediaSessionDelegate {
On 2016/02/26 at 03:22:50, DaleCurtis wrote:
> Ditto on the export. Surprised your browser tests compile.

I don't actually use MediaSessionDelegateAndroid in browser tests. It is used in
the integration tests (Android's instrumentation tests) but it doesn't access
the class directly so I don't think CONTENT_EXPORT is required.

[commit-bot: I haz the power, 2016-02-26 16:18:14]

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

[commit-bot: I haz the power, 2016-02-26 16:18:15]

Dry run: Try jobs failed on following builders:
  linux_android_rel_ng on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/linux_androi...)

[whywhat, 2016-03-01 03:12:10]

lgtm with nits

https://codereview.chromium.org/1698933004/diff/220001/content/browser/BUILD.gn
File content/browser/BUILD.gn (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/BUILD....
content/browser/BUILD.gn:353: "media/session/media_session_delegate_default.cc",
nit: media_session_delegate_impl.cc?
Seems like the pattern is to have Impl and Android so far :)

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
File content/browser/media/media_web_contents_observer.cc (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/media_web_contents_observer.cc:41:
session_controllers_manager_.Clear(render_frame_host);
nit: s/Clear/RenderFrameDeleted?

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/media_web_contents_observer.cc:101: const MediaPlayerId
id(render_frame_host, delegate_id);
nit: s/id/player_id to distinguish with other ids?

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
File content/browser/media/media_web_contents_observer.h (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/media_web_contents_observer.h:16: #include
"content/browser/media/session/media_session_controllers_manager.h"
nit: would it hurt readability if we dropped _controllers_ from the name?

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
File content/browser/media/session/media_session.cc (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/session/media_session.cc:198: return audio_focus_state_ ==
State::SUSPENDED;
Given there's three states, your comment in the header about this function
doesn't seem to be correct.

[mlamouri (slow - plz ping), 2016-03-07 19:28:09]

https://codereview.chromium.org/1698933004/diff/220001/content/browser/BUILD.gn
File content/browser/BUILD.gn (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/BUILD....
content/browser/BUILD.gn:353: "media/session/media_session_delegate_default.cc",
On 2016/03/01 at 03:12:10, whywhat wrote:
> nit: media_session_delegate_impl.cc?
> Seems like the pattern is to have Impl and Android so far :)

I wouldn't say that WMPI/WMPA is a common pattern. For system-specific
implementation, my understanding is that _default.cc _{platform}.cc is more
common.

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
File content/browser/media/media_web_contents_observer.cc (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/media_web_contents_observer.cc:41:
session_controllers_manager_.Clear(render_frame_host);
On 2016/03/01 at 03:12:10, whywhat wrote:
> nit: s/Clear/RenderFrameDeleted?

Done.

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/media_web_contents_observer.cc:101: const MediaPlayerId
id(render_frame_host, delegate_id);
On 2016/03/01 at 03:12:10, whywhat wrote:
> nit: s/id/player_id to distinguish with other ids?

Done.

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
File content/browser/media/media_web_contents_observer.h (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/media_web_contents_observer.h:16: #include
"content/browser/media/session/media_session_controllers_manager.h"
On 2016/03/01 at 03:12:10, whywhat wrote:
> nit: would it hurt readability if we dropped _controllers_ from the name?

It wouldn't hurt readability but I think it wouldn't be entirely correct. As
Dale suggested, we might need to consolidate some classes here but let's worry
about that in a follow-up?

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
File content/browser/media/session/media_session.cc (right):

https://codereview.chromium.org/1698933004/diff/220001/content/browser/media/...
content/browser/media/session/media_session.cc:198: return audio_focus_state_ ==
State::SUSPENDED;
On 2016/03/01 at 03:12:10, whywhat wrote:
> Given there's three states, your comment in the header about this function
doesn't seem to be correct.

My comment says that ::IsSuspend() (not this method) should be `==
State::SUSPENDED` instead of `!= State::ACTIVE` so I created this method in
order to not have to refactor usage of ::IsSuspended() actually checking for
`!IsActive()`.

I rephrased my comment a bit to make it clearer.

[mlamouri (slow - plz ping), 2016-03-12 01:32:17]

The CQ bit was checked by mlamouri@chromium.org

[mlamouri (slow - plz ping), 2016-03-12 01:32:18]

The patchset sent to the CQ was uploaded after l-g-t-m from jochen@chromium.org,
dalecurtis@chromium.org, avayvod@chromium.org
Link to the patchset: https://codereview.chromium.org/1698933004/#ps300001
(title: "rebase")

[commit-bot: I haz the power, 2016-03-12 01:32:59]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1698933004/300001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1698933004/300001

[commit-bot: I haz the power, 2016-03-12 04:48:01]

Committed patchset #16 (id:300001)

[commit-bot: I haz the power, 2016-03-12 04:49:29]

Description was changed from

==========
Make MediaSession a runtime-enabled feature on Desktop.

At the moment, it's only compiled on Android, this CL compiles it on all
platforms
but it's only enabled by default on Android. Other platforms have to pass
--enable-default-media-session flag to the command line.

This CL move all the media session related files from
content/browser/media/android/
to content/browser/media/session/ and tear down the Android specific code from
MediaSession into a delegate. The Desktop delegate simply pauses other tabs when
a
media session gets activated in a tab.

It is also moving the MediaSessionController handling to a separate class so it
can move from MediaWebContentsObserverAndroid to MediaWebContentsObserver
easily.

BUG=587471
==========

to

==========
Make MediaSession a runtime-enabled feature on Desktop.

At the moment, it's only compiled on Android, this CL compiles it on all
platforms
but it's only enabled by default on Android. Other platforms have to pass
--enable-default-media-session flag to the command line.

This CL move all the media session related files from
content/browser/media/android/
to content/browser/media/session/ and tear down the Android specific code from
MediaSession into a delegate. The Desktop delegate simply pauses other tabs when
a
media session gets activated in a tab.

It is also moving the MediaSessionController handling to a separate class so it
can move from MediaWebContentsObserverAndroid to MediaWebContentsObserver
easily.

BUG=587471

Committed: https://crrev.com/91873409e26ae1547dcd759735348884823f66ad
Cr-Commit-Position: refs/heads/master@{#380850}
==========

[commit-bot: I haz the power, 2016-03-12 04:49:31]

Patchset 16 (id:??) landed as
https://crrev.com/91873409e26ae1547dcd759735348884823f66ad
Cr-Commit-Position: refs/heads/master@{#380850}