Notes, May 28

On 05/27/2015 11:01 PM, Harald Alvestrand wrote:
> Overarching issues:
>
> Work style - does the current style work for us?
> In what roles would more manpower help get things resolved faster?
>
> Mediacapture-main:
>
> Pull requests:
> -------------
> #170 Errors - not ready yet?
Still working on errors. Worried by low response to the "here's the
state" message.
Stefan will send out a message saying "We'll assume that this is the
plan if we don't hear anything else by Tuesday".

> #171 HTML 5.1 reference for srcObject - ready?
Dan has found issues with the PR (media load algorithm). There's a huge
bunch of steps in HTML 5.1 that we need to check to see if they matter
to us - there were some specific exceptions in the original text, we
need to check this against HTML 5.1 objects.
Adam will review. Perhaps Justin also should be asked to review?

Harald will send a poke to the list asking for review.

> #172 Use [Exposed] - ready?

Merged.

New PR: #175 - "addtrack is not used by this spec, but used by others".
Adam will call attention to it in the "addtrack signal" thread.

Note from Adam: We have weird text on event handlers - we need to
separate it into "event is fired" text and "eventhandler exists" text.
The DOM event handling system is more complex than this text seems to
suggest, so the text is misleading. Adam to propose changes.

>
> Travis is still failing because of the link problem on these - not rerun
> recently?
>
> Issues:
> -------
> #118 Practical algorithm - HTA was to push this to list.
> #127 Timeout - still sleeping
> #157 rename enumerateDevices - close due to LC disposition?
Closing with comment.
> #161 Remove Direct Assignment - see #171
> #162 MediaStreamError - working on it, see #170
> #163 Use [Exposed] - see #172
> #173 Consider specifying a minimum length for deviceId - NEW
List discussion needed. Harald will poke.

New: #174 - synchronous parts of getUserMedia - likely not, but there is
an issue here.
>
> LC issues:
> ----------
> Stuff that needs editing in rather than more discussion:
> - LC-3022 Latency constraint - Peter's promised a PR
> - LC-3027 Internationalization of "message" - needs some words
Adam will craft some words.
> - LC-3019 IDL errors - we *can* patch this to make IDL checkers happy
> (more dummy types). Should we?
Suggest asking Jan-Ivar to make a PR. #176 filed.
>
>
> Webrtc-PC:
>
> Pull requests:
>
> #29 mediadiscarded - time to find another Martin?
Martin would like someone else to take over. Ask Cullen?
Dan will ping Cullen and Justin about it (the main opinion-havers).

> #195 replaceTrack - the first cut didn't cut it, but list discussion
> seems to suggest that we now know approx. what we want.
We are really close to what J-I has in the PR - Stefan can ask J-I to
update to reflect present consensus.

> #215 Expiring certificates - merge?

Dan offers to merge it without a whitespace change.

> #221 Define when RTCRtpReceivers are created and dispatched - Adam?

Merge-ready. "Media description" is right. Adam will rebase and merge.
> #222 Replace operations array with a chain - does anyone care?
> #225 Support for pending and current SDP - still discussion with
> Jan-Ivar in JSEP
> #228 QoS API (priority) - debated hotly on list. Waiting for
> Martin/Peter to spec alternative (RTPSender.EncodingParameters, with a
> get/set model for changing).
> #229 Removing optionality from constructors - good?

List discussion needed - it's reasonable for most things, but Martin has
a comment.

#231 - merged!

#233 - "bufferlow" event - let's wait, this is not a simple topic.

We did not get to bugs.

>
> Some selected bugs:
>
> #129 Encoding Parameters -see discussion on #228
> #140 DataChannel.onerror() needs an error parameter specified (another
> Error! Yay!)
> #151 close is synchronous and idempotent
> #179 IP address revelation - where do we think we are?
>
>
>
>
>
>


-- 
Surveillance is pervasive. Go Dark.

Received on Thursday, 28 May 2015 15:07:56 UTC