Re: Agenda, August 27 (V2)

Den 27. aug. 2015 15:23, skrev Harald Alvestrand:
> Apologies for not getting to this yesterday.
> 
> Overarching items:
> 
> I'd like to chat about constraint registrations. Only good thing that
> came out of the flicker discussion is that people are reminded it's not
> settled.

We don't seem to have a bug for it (any more?)
#282 filed.

> 
> The Accessible Media discussion (the call with Nigel) - thoughts to
> share for Dan and Stefan to bring to the meeting?

We're not likely to do what Nigel suggests (at least not in this version).
"Please start work on an extension spec" seems to be the mantra.

> Another WebIDL misfeature on returning dictionaries has surfaced - brief
> orientation. (Bug is webrtc-pc:275)

Unhappiness noted. Perhaps TPAC topic?

> Media Capture and Streams:
> 
> Pull requests:
> 
> #211 Device capabilities - good enough to merge (with "info.capabilities")?

Remove sentence about output in description and merge?
Stefan will take on.

> #226 Typedef MediaStreamError as Object - Dan needs to check if
> reference is OK.

#235 Remove static like definitions of attributes and methods

This is in response to #233. This seems to trigger a respec bug, where
<dfn id=xx>aa</dfn> twice will barf, even if the "xx" is different, as
long as the "aa" is the same.

The same bug happens with the attribute called "state" in a webrtc-pc thing.
Adam will consult with Dom and file a Respec bug if needed.

Some more fixes needed, so will have another commit before merging.

> 
> Issues:
> 
> #118 Practical algorithm
> #127 Timeout/cancellation - still haven't defined the icebox for "not
> this round". Suggest to close, possibly with a "for future study" tag.
> #189 getUserMedia TypeError

Anne agrees. Let's do it. Adam will write a PR.

> #193 Adaptive frame rate
> #198 Device capabilities - see #211
> #212 Privacy/security review - Peter?

Waiting for Peter.

> #221 Unconnected MSTracks - seems to be going away based on createRTPSender.

Closed with reference to webrtc-pc.#271

> #222 Groupid is origin-unique - answer is yes, who edits?

Peter has promised text.

> #225 Track vs sink section 5 - Dan
> #227 Overconstrained event handler attribute description - Dan
> #230 WebIDL reference - no change
> #232 Clone a track - Adam?

Other specifications use attributes as values. Anne's approach with
"internal slots" all over the place seems pedantic. Lots of text work,
but no real changes.
Can we find someone else to create the PR?
Don't prioritize at the moment.

> #233 MediaStream.active and friends - Adam?

See PR #235.

> #234 PowerNetworkFrequency - I've concluded it's a bad idea, just have
> to convince
>    some other Googlers...
> 
> WebRTC-PC:
> 
> Pull requests: 20 (most of them by pthatcher)
> 
> Need to decide what's mergeable now, and what should wait for the interim.
> 
> #236 Replace operations array - adam?

Closing. No consensus from WG.

> #237 replaceTrack - for F2F?
> #240 Move certificate management - dan? Or adam?
> #241 *.transport attributes - seems to be little/no resistance, and it's
> prev-meeting-compatible - merge?

Adam has sent a PR on Peter's PR to fix a respec error.
Adam will merge when Peter accepts the PR.

> #242 SyntaxError on malformed ICE candidate - Stefan?
> #254 MaxBitrate - obsoleted by #268?
Closed.

> #255 getCapabilities - obsoleted by #269? Close this?
> #256 Fix prose on getStats w/o selector
> #258 Codec stuff - meeting?
> #261 Ref for ICE Completed - Peter?

Back to Cullen.

> #266 Tidy setLocal/setRemote - Adam

There is some change to the throwing of errors in the resolution (state
checks).
That should be a separate issue. Merge and raise it as an issue. Adam.

> #268 maxBitrate v2

Merged.

> #269 getCapabilities v2

Leave open. Discussion was in July.

> #270 SCTP Transport objects
> #271 createRTPSender
> #272 RTPSenderInit (alternative to createRTPSender)
> #273 Bunch of fields - suggest aiming to leave FEC and simulcast out of
> 1.0, but that's likely a F2F topic.
> #279 CreateRTPReceiver
> #280 more ICETransport attributes and methods
> #281 Caution people to read through the entire spec before implementing
> (partial RTCPeerConnection) - suggest to merge

Merged.

> 
> Bugs
> 
> #275 is the only one I can think of offhand.

Discussed above.

#276 Client-to-mixer

Seems OK to always send if negotiated. IETF problem where to place it.
Leave open for a few days, then close as "no change".

> 
> 
> 
> 
> Issues:
> 
>

Received on Thursday, 27 August 2015 15:00:28 UTC