Re: Agenda, Jan 26

Den 26. jan. 2017 14:06, skrev Harald Alvestrand:
> Up front stuff:
> 
> - Chair changes for Q2/Q3

Harald announced his sabbatical (post-Easter)
> 
> - WG interim decisions review & PR writing assignments
> 
> Mediacapture-main
> =================
> Pull requests (glorious count!)
> -------------
> #422 Clarify when and where devicechange events fire. ()

Missing comment responses, Travis barfs. Restarted
> 
> Issues
> ------
> #380 Define restrictions on device-info permission. (alvestrand)
> #384 Should we use [EnforceRange] on min/max in constraints? (aboba)
> #385 ConstrainLong only concern positive values (aboba)
> #386 What does "getSettings" return? (alvestrand)
> #396 Explanation of "advanced" constraint sets is confusing (burnburn)
> #400 Add text to extensibility for consumer of MST (burnburn)
> #404 Revive createObjectURL? (stefhak)
> Can be closed after interim decision review

Stefan will write up the closing text and close the issue.

> 
> #409 Clarify behavior of using unrelated constraints (alvestrand)
> #414 Devicechange events when not focus - permitted or forbidden (jan-ivar)
> I think this is closed
> 
> #425 Do we update legacy methods to keep up with the spec? (burnburn)
> #428 Implementation Suggestions section needs review (stefhak)
> #429 Specify that getSettings omits non-applicable settings (burnburn)
> 
> WebRTC-PC
> =========
> Pull requests
> -------------
> #624 Upscale allowed (fluffy)
Stefan will ping Cullen offline.

> #776 Support for OAuth in TURN credentials (Issue 714 patch) (stefhak)
Needs more work to match the option selected.
WG meeting - misi mentioned "realm" issue, but WG decided not to tackle.

> #971 Clarify wording on TypeError from addIceCandidate. (aboba)
Not fixed yet.

> #988 Add RTCOfferOptions.reofferOptions. ()
Closed based on WG consensus.

> #990 Add an explicit stats selection algorithm. ()
Submitter needs to conform to WG consensus. Jan-Ivar has promised input.

> #994 Changing "non-null" to "missing" to match IDL terminology. ()
Merged.

> #996 Describe when an RTCSctpTransport is created/set to null. ()
It's updated with WG consensus. Merged.

> #999 transceiver.stop() sends a BYE ()
Merged.

> #1000 Separated auth dictionaries for STUN/TURN (issue 714) ()
> Can be closed given that WG went for "hybrid"
Needs more work to match the option selected.

> #1001 Effect of a BYE on RtpReceiver.track ()
Needs review.

> #1002 Event when a transceiver is stopped via remote action ()
Merged.

> #1003 Change SetLocalDescription to require unchanged offer/answe ()
Needs review. Asking Taylor.

> #1004 currentRemoteDescription.sdp need not match setRemoteDescri ()
Update to say "and the local"

> #1005 Add offerToReceive* as legacy extensions ()
Too fresh to merge.

> 
> Issues
> ------

The requirement for CR is "we believe it's feature complete".

> I got a list of issues from Dom where he suggested that we could call
> out some issues
> as "really needing work", while others were "we know what to do".
> I've called out those "really needing work" with a star below.
> 
> #295 * Guidance for extending objects vs extending Stats needed (alvestrand)
> #305 Describe what happens when media changes (fluffy)
> #526 NetworkError event is not defined and might not be needed (aboba)
> #578 Need to specify precisely when MID generation happens (aboba)
> #661 Add informative table of all things that can cause negotiat (burnburn)
> #671 * Processing remote MediaStreamTracks without MediaStreams in
> (alvestrand)

MSID patches are in flight. Need commentary!

> #687 Clarify reasoning behind and mitigation of privacy issues (
> (dontcallmedom)
> #688 Indicators of usage and data flow (PING review) (dontcallmedom)
> #690 Information available prior to permission prompt (PING revi
> (dontcallmedom)
> #709 offerToReceiveAudio/offerToReceiveVideo remain in implement (aboba)
> #714 STUN/TURN OAuth token auth parameter passing (aboba)
> #729 RTCStats timestamp source ambiguous (alvestrand)
> #763 * Handling of simulcast errors (aboba)

Seems to be OBE - we're just rejecting the setParameters, no additional
machine-parsable data.

> #787 Integrate RTCRtpTransceiver into set local/remote steps ()
> #788 Clean up remaining uses of 'set of receivers' ()
> #795 * Specify when a data channel's ID is assigned, and what the
> (taylor-b)

Closed due to lack of memory of what needed fixing.

> #825 * Need IDP Invalid Token  ()
> #826 * Need IDP Expired Token error  ()
> #830 * Need Custom Error for IdP  ()
> #831 * Define how long should the IdP timeout timer should be
> (martinthomson)

We need a block proposal here. Either come up with a proposal, or cut
IDP from the spec. When (if) this is specified, will Firefox implement it?
Stefan to ping Cullen together with #624.

> #842 Define JSEP terms like "bundle" and "bundle-policy" in term (burnburn)
> #845 "Throw a FooError" steps not written in a consistent manner (aboba)
> #849 Specify an AllowUnverifiedMedia RTCConfiguration property  (fluffy)
> #857 The RTCPeerConnectionIceErrorEvent constructor should have  (aboba)
> #881 Align getAlgorithm return value with Web Crypto (foolip)
> #921 currentRemoteDescription.sdp -- does it need to match the l (aboba)
> #930 Ambiguous wording in addIceCandidate (taylor-b)
> #932 Should IDP Login error cary the idpLoginUrl  (martinthomson)
> #942 Meta: auto-publish changes to the spec (alvestrand)
> #950 Change SetLocalDescription to require unchanged offer/answe (aboba)
> #954 Receive a track multiple times (stefhak)
> #959 Advanced Peer-to-peer Example (aboba)
> #961 Effect of a BYE on RtpReceiver.track (aboba)
> #962 Event when a transceiver is stopped via remote action (aboba)
> #964 * Handing SDP with more than one identity  (martinthomson)

This sparked a discussion on what to do with issues that resolve slowly.
Deadline-driven work from the chairs seems to be needed.

> #979 When is an RTCSctpTransport created and destroyed? ()
> #997 * RTCRtcpMuxPolicy of "negotiate" should be marked as an "at- ()

It's a "non-surprise" marker - if nobody does, it will not block CR-PR
transition.

Note: We may consider marking IDP as "at-risk". It may be a necessary
impetus to either getting it implemented or getting another solution on
the table.
Note from Bernard: People experimenting with PERC are also dependent on
IDP to have a complete story.

> 
> If we know what to do on all the others, it's just a matter of doing it.
>

Received on Thursday, 26 January 2017 16:03:24 UTC