search

fluffy / webrtc-w3c

W3C Specification for WebRTC
30 stars 18 forks source link

Fixing peerIdentity, removing noaccess #10

Closed martinthomson closed 10 years ago

martinthomson commented 10 years ago

I made some tweaks to the peerIdentity constraint, removed the much maligned noaccess. I've also moved peerIdentity from being a constraint to a property of the gUM "constraints" dictionary.

There's extra text here explaining what to do when peerIdentity constraints get mixed up. There's more to do on this, but that belongs in WebRTC.

alvestrand commented 10 years ago

On 02/06/2014 01:49 AM, Martin Thomson wrote:

I made some tweaks to the peerIdentity constraint, removed the much maligned noaccess. I've also moved peerIdentity from being a constraint to a property of the gUM "constraints" dictionary.

There's extra text here explaining what to do when peerIdentity constraints get mixed up. There's more to do on this, but that belongs in WebRTC.

At a glance, this looks good to me. I'm happy with making the spec simpler, and only specifying what we know we need now.

One thing that might be worth mentioning: When you track.clone() a track, the peerIdentity constraint is inherited. I think it's obvious from the spec of constraints that it does, but it might be worth calling out.

fluffy commented 10 years ago

Can you add text emphasizing what happens on clone and check that all the WebIDL got updated ?

martinthomson commented 10 years ago

@fluffy, the clone text is in (I have commented out text regarding stream processing, but I think that should remain commented out). The WebIDL is good as far as I'm concerned.