Refactor content::RenderAudioSourceProvider as webkit_media::WebAudioSourceProviderImpl.

Refactor content::RenderAudioSourceProvider as webkit_media::WebAudioSourceProviderImpl.

Doing so prevents the silliness of passing the same pointer into WebMediaPlayerParams twice. Furthermore, the switch-checking logic and render view ID code is better contained as part of RenderViewImpl as opposed to being stuffed inside RenderAudioSourceProvider's constructor.

BUG=136442
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=175948

[scherkus (not reviewing), 2013-01-08 22:00:57]

Got tired of dealing w/ the duplicate AudioSourceProvider and AudioRendererSink
pointers and the resulting lifetime management issues

miu: I read through your doc and I believe this CL won't hamper your plans

[miu, 2013-01-09 18:48:35]

lgtm

https://codereview.chromium.org/11830004/diff/2013/webkit/media/webaudiosourc...
File webkit/media/webaudiosourceprovider_impl.cc (right):

https://codereview.chromium.org/11830004/diff/2013/webkit/media/webaudiosourc...
webkit/media/webaudiosourceprovider_impl.cc:84: is_running_ = true;
Weirdness in existing code (not part of your change, but...): Perhaps you should
remove this statement here, since calling sink_->Start() doesn't mean the sink
is consuming audio yet.

https://codereview.chromium.org/11830004/diff/2013/webkit/media/webaudiosourc...
File webkit/media/webaudiosourceprovider_impl.h (right):

https://codereview.chromium.org/11830004/diff/2013/webkit/media/webaudiosourc...
webkit/media/webaudiosourceprovider_impl.h:72:
scoped_refptr<media::AudioRendererSink> sink_;
nit: Since sink_ is set via the ctor initializer list and never changed anywhere
else in the class, consider making the type const scoped_refptr<...> here.

[DaleCurtis, 2013-01-09 21:51:38]

lgtm

[scherkus (not reviewing), 2013-01-09 23:16:06]

https://codereview.chromium.org/11830004/diff/2013/webkit/media/webaudiosourc...
File webkit/media/webaudiosourceprovider_impl.cc (right):

https://codereview.chromium.org/11830004/diff/2013/webkit/media/webaudiosourc...
webkit/media/webaudiosourceprovider_impl.cc:84: is_running_ = true;
On 2013/01/09 18:48:35, Yuri wrote:
> Weirdness in existing code (not part of your change, but...): Perhaps you
should
> remove this statement here, since calling sink_->Start() doesn't mean the sink
> is consuming audio yet.

Discussed offline, ARS::Start() does indeed start playback! Play() is for after
calling Pause()