- From: Adam Bergkvist <adam.bergkvist@ericsson.com>
- Date: Wed, 12 Jun 2013 12:54:00 +0200
- To: robert@ocallahan.org
- CC: Jim Barnett <Jim.Barnett@genesyslab.com>, Travis Leithead <travis.leithead@microsoft.com>, "public-media-capture@w3.org" <public-media-capture@w3.org>
On 2013-06-12 11:38, Robert O'Callahan wrote: > On Wed, Jun 12, 2013 at 9:03 PM, Adam Bergkvist > <adam.bergkvist@ericsson.com <mailto:adam.bergkvist@ericsson.com>> wrote: > > I agree that a final ended state is useful, but it would make more > sense to base it on something else that all tracks being removed or > becoming ended. I'll take a stream received via a PeerConnection as > an example. > > dispatched at B-side by the addstream event: active > a-side removes all tracks: inactive > a-side re-adds live tracks: active > a-side ends-all tracks: inactive > a-side does removeStream(): ended > > > What if the a-side's stream reaches the permanently-ended state? > > Perhaps only streams created by getUserMedia() and PeerConnection > should reach the ended state. That may bring us back to [1]. > > > OK, we could have a separate permanently-ended state, and we could > define that state separately for each kind of stream. We could, then, > give constructed MediaStreams the sort of API I proposed, and/or some > way to explicitly end the stream. > > But if we do those things, I'm not sure why an author would want to use > "inactive" instead of just "ended". > Perhaps it's best to keep it as simple as possible. We have moved most of the stuff from MediaStream to MediaStreamTrack and the stream state is, after all, a convenience thing, so it might be a bad idea to over complicate it. I think Jim's active/inactive proposal is enough. Other information is mostly given by the application context; for example in my state sequence above, the removestream event tells that this inactive stream was removed on the a-side and can be considered permanently ended. /Adam
Received on Wednesday, 12 June 2013 10:54:24 UTC