Bug 145806 – Implement MediaSessionManager to keep track of all MediaSessions

• Status: RESOLVED
• Resolution: FIXED
• Component: Media

[Matt Rajca]

As described in the Media Session spec, MediaPlayPause events need to be handled by each Media Session, and we need a MediaSessionManager to keep track of them.

[Radar WebKit Bug Importer]

<rdar://problem/21302849>

[Matt Rajca]

Created attachment 254580 [details]
Patch

[Darin Adler]

Comment on attachment 254580 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=254580&action=review

> Source/WebCore/Modules/mediasession/MediaSessionManager.cpp:49
> +    m_sessions.remove(m_sessions.find(&session));

Normally it’s not good to use remove/find. I’d suggest using a HashSet or ListHashSet for this collection instead of a vector.

[Matt Rajca]

Created attachment 254585 [details]
Patch

[Matt Rajca]

(In reply to comment #3)
> Comment on attachment 254580 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=254580&action=review
> 
> > Source/WebCore/Modules/mediasession/MediaSessionManager.cpp:49
> > +    m_sessions.remove(m_sessions.find(&session));
> 
> Normally it’s not good to use remove/find. I’d suggest using a HashSet or
> ListHashSet for this collection instead of a vector.

Settled on HashSet since we don't care about the ordering of MediaSessions. Thanks!

[Eric Carlson]

Comment on attachment 254585 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=254585&action=review

> Source/WebCore/Modules/mediasession/MediaSessionManager.h:41
> +    const HashSet<MediaSession*>& sessions() const { return m_sessions; }

Is there any reason to not make this private?

[Matt Rajca]

(In reply to comment #6)
> Comment on attachment 254585 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=254585&action=review
> 
> > Source/WebCore/Modules/mediasession/MediaSessionManager.h:41
> > +    const HashSet<MediaSession*>& sessions() const { return m_sessions; }
> 
> Is there any reason to not make this private?

I have other uncommitted code here locally that requires it to be public, but that is probably going to get refactored anyway. I'll remove the accessor for now.

[Matt Rajca]

Created attachment 254599 [details]
Patch

[WebKit Commit Bot]

Comment on attachment 254599 [details]
Patch

Clearing flags on attachment: 254599

Committed r185389: <http://trac.webkit.org/changeset/185389>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.