- From: Randell Jesup <randell-ietf@jesup.org>
- Date: Sun, 05 Jan 2014 12:40:38 -0500
- To: public-webrtc@w3.org
On 1/2/2014 10:42 AM, Iñaki Baz Castillo wrote: > Hi Stefan, thanks for your response. Comments inline: > > 2014/1/1 Stefan Håkansson LK <stefan.lk.hakansson@ericsson.com>: >>> What is wrong with a MediaStream.stop() method that calls >>> stop on every track? >> I think it is mostly about clearness. If the same track is part of more >> than one MediaStream, and you call stop on one of them - should the >> track stop in the other one(s) or not? > What I mean is a MS.stop method that iterates over all the MST and > call stop() on each. So if a track also belongs ot other MS then the > behaviour would be the same as if the user call stop() on the given > MS. I see no problem, MS.stop is just a convenient method. I agree. All that leaving it out of MediaStream means is that every framework will implement their own version of it, and if a framework has a bug dealing with changes to the underlying browser implementations (new track types, etc), that bug will be painful to fix since there's a huge problem with frameworks getting fixed in stone by webpages/apps (the never-update problem). Obvious convenience functions like this both make it easier to use, especially for non-experts looking to make simple use of WebRTC/MediaStreams/gUM, and reduce the chance of application bugs. THey also fit in with the idea of providing mid/higher level simple-to-use functionality, and the proposed lower-level nuts-and-bolts expert-level access. IMHO. > >> I think the design has been iterated many times (last time was during >> the f2f meeting at TPAC), and most people are happy as it is. It is not >> that hard to loop over the tracks in a MediaStream! > Probably not, but what about if MS has tracks other than audio and > video tracks (i.e. text tracks)? Maybe this does not make sense but > let's see what happen in the future. I would not like that today's > applications become broken in the future > because MS get new kind of tracks and the today's code just call > stop() on audio and video tracks. Right. >> I think that if we should add anything, then it would be a stop/return >> on the source level - sort of allowing the app to return the right to >> use a source in a more direct way than having to stop all tracks the >> source serves. But so far there has not been a strong desire for this >> feature. > How to access the "source level"? Currently what we get via gUM is a > MS, the source is implicit and not addressable. This is similar to the issues surrounding the 'doohickey'... -- Randell Jesup -- rjesup a t mozilla d o t com
Received on Sunday, 5 January 2014 17:41:46 UTC