Re: replaceTrack proposal

On 4/14/15 4:56 PM, Martin Thomson wrote:
> If setTrack() was an acceptable name, then we would be better off with:
>
>    sender.track = theReplacement;

sender.track is one of the proposals. 
https://github.com/w3c/webrtc-pc/pull/196

By extension I can't see anything logically wrong with setTrack as a 
name for the method version. Once we add more properties to RtpSender, 
the "replace" language may not carry over well.

> As I understand it, replaceTrack can fail, asynchronously, if the
> underlying track is incapable of producing a compatible RTP stream.
> That might not be the case in current implementations, but those
> implementations are probably not using encoding cameras that can't
> change codecs.

Is this not deterministic? For comparison, where would one learn of such 
failures from tracks added with pc.addTrack()?

.: Jan-Ivar :.

Received on Wednesday, 15 April 2015 06:28:08 UTC