Notes, March 10 2016

Den 10. mars 2016 14:24, skrev Harald Alvestrand:
> The usual:
> - Echidna publishing is awaiting an explicit WG declaration of consensus.
> And the respec bugs!
> Travis is known to be broken. Ignore it.
> Mediacapture-main
> =================
> We're closing in on CR call for consensus!
> Pulls
> -----
> #319 Text incorporating the Permissions API's definition of permi
> (martinthomson)
> Good review from Adam. Still no response from Martin or Mozilla.

Waiting for update.

> #321 Removing track attributes "remote" and "readonly" (see Issue ()
> Awaiting consensus declaration?

Let's give it another week.
> #323 Remove unused images ()
> Simple? We can always pull them out of history if we need them.


> Issues
> ------
> #305 Use permissions API model for describing permissions? (alvestrand)
> See PR #319.
> #316 MediaTrackSettings/ConstraintSet/Capabilities... should link (burnburn)

No PR, and no Dan.
> #317 Switch getUserMedia permission-rejection to NotAllowedError
> (alvestrand)
> No news here. Terminology is hard.

We conclude that we can go with NotAllowedError. Adam will make PR.

> #320 Should we remove track attributes "remote" and "readonly"? ()
> I think the chairs need to call consensus on this one; would prefer to
> remove.


> #322 Browsing sessions and device IDs ()
> Sounds like consensus for no change to me, but not 100% clear yet.

Dom suggested we might want to remove a section. Not clear if that's
what the Mozillans want too.
Stefan will follow up.

> =========
> Pulls
> -----
> #463 Add more explicit language about transceivers, MIDs and roll (jan-ivar)
> Wrong assignment from last time. Adam's comments not replied to. Assign
> pthatcher.

Adam merged it. Adam thinks a lot more text is needed; it's not clear now.
We have senders without a track, but track property of sender is not
More info properties are not happy - can't be initialized with no track.
Action Adam: File issues.

> #466 Use an enum to describe directionality of RTP Stream (adam-be)
> #467 Use enum for voiceActivityDetection (adam-be)
> #471 Use enum for RTCDataChannel’s ordered attribute (adam-be)
> Enum group.

Discussion has died down :-(
Action: Suggest assign as next interim topic, Harald to write up a
position summary.

> #494 proposed fix to issue #492 (adam-be)
> Can we close this now?

Closed with comments.

> #519 Relation between local track and outgoing encoding (Issue #4 (aboba)
> Bernard, are you OK with this?


> #520 Add text about 'remote sources' and how they are stopped (is (aboba)
> If Bernard's OK, let's merge.


> #527 Enable trickling of end-of-candidates through addIceCandidat (taylor-b)
> Comment not addressed.

Needs response from Taylor. Pinging.

> #533 Add RTCIceTransportState Enum (taylor-b)
> Comments not addressed. This isn't yet used in a description of how the
> ICEAgentState is derived.

As above.
> #535 update MID to be random values when not received in offer ()
> Just merge?

Waiting for response from Cullen.

> #538 [WIP] Trigger echidna automatic publications using Travis
> (dontcallmedom)
> Admin. Not ready.
> #539 Update list of sections from JSEP v13 ()
> Admin. Merge.


#540 Removal of "Server cannot be reached" errors
Robin Raymond suggested that we shouldn't call this out as an error,
because flaky connectivity or IPv6 candidates on isolated networks could
give lots of them.

Suggestion is that we don't return this event, return it only if no
candidates can be reached, or make it a separate event.

Closed PR, Bernard will make an issue out of it

> Issues
> ------
> #142 "Hold" unspecified (aboba)

Waiting for the enum PRs to be resolved.

> #227 .mid attribute - random or Null when MID is not present in S (fluffy)

PR exists.

> #230 Add support for WebRTC Data Channel in Workers ()

Closed as LATER.

> #253 Assurance that requests to IdP proxy originate from the user
> (martinthomson)

On Martin's plate. December update was "no time found yet".

> #257 ICE Candidate should have accessors for protocol-relevant it
> (alvestrand)

Ping delivered. Might be solved.

> #295 Guidance for extending objects vs extending Stats needed (alvestrand)

Ping delivered.

> #296 Debugging ICE problems needs more info (aboba)

Bernard knows what to do.

> #305 Describe what happens when media changes (fluffy)


> #336 No definition of "applying an ICE candidate" (adam-be)

Partly solved by #507. Adam will raise a new issue for remaining stuff.
(Note for adam-be: Please check last week's notes for other issues you
may have promised to write!)

> #337 Interfacing between WebRTC spec and JSEP (burnburn)

No dan.

> #363 m-line match up doesn't work with RtpSender.mid or RtpTransc
> (pthatcherg)

Fixed, closing.

> #364 Browser calls close() on failed rollback (pthatcherg)

Reassigned to Adam.

> #369 addTrack's streams parameter is unused. (adam-be)

Linked to the transceiver stuff. Not written up yet.

> #375 true as default values for dictionary is bad practice (adam-be)

See enum cluster

> #389 Should have a "closed" RTCPeerConnectionState (pthatcherg)

Kind of duplicate of #530. Moving the state is reasonable.
Note: We need to add text that fires events at the
onconnectionstatechange handler - currently that's missing.

Adam will file a new bug for firing connectionstatechange events.

> #413 Describe what happens to transceivers at rollback (pthatcherg)
> #440 Use interfaces not dictionaries (adam-be)
> #442 Impossible to know if ICE agent is "finished checking", for  (taylor-b)
> #457 Non-normative ICE state transition diagram (aboba)
> #478 Clarify relation between a local track's settings and how it (stefhak)
> PR #519
> #483 Signaling a=end-of-candidates (aboba)
> WG has consensus on "yes".
> #492 avoid risk of reentrancy (ShijunS)
> #495 File a WebIDL issue to add the DOMException names WebRTC nee (burnburn)
> #496 Certificate API: Do we need certificate introspection? (aboba)
Fixed, merged.

> #513 Handling the stopping of a track should be described (adam-be)
> #517 Clarify ICE consent freshness feedback (aboba)
> #518 PING review of webrtc-pc spec ()
> #521 RTCPeerConnectionIceErrorEvent errorCode not defined (aboba)
Fixed, merged

> #522 Setting iceCandidatePoolSize to a non-zero value may initiat (adam-be)
> #523 Add API to control encrypted header extensions for RTP ()
> #526 NetworkError event is not defined and might not be needed (adam-be)
> #529 MediaStreamTrack clarifications needed(?) (stefhak)
> #530 RTCPeerConnectionState is a better home for "closed" than RT (adam-be)
> #534 Should remote tracks have .readonly false or true? ()
> Hopefully obsoleted by removing .readonly

Received on Thursday, 10 March 2016 16:03:21 UTC