See also: IRC log
bernard: we're trying
to tackle issue clusters in this meeting
... we're still
interested in knowing features not on the radar for
implementation
<dom> PR 1430 to address Issue 1383
bernard: PR1430 seems to resolve cullens concerns
bernard: We lost ekr, so we move on and come back to PR 1430 later
PR 1430 resolution: ekr will comment on the PR
Issue 1174: The issue can make it hard to implement WebRTC on top of ORTC
Issue 1174: Resolution is for Taylor to do a PR to remove the odd behavior
Issue 1178: When are *Transport objects created? Resolution: the group agrees with Taylor's suggestion
Taylor will create an PR
bernard: We need to continue to discuss Issue 1406 on github.
peter: Issue is if a new IceTransport is created or not
<mt_____> The problem is whether we make new IceTransport objects, or we mutate the same object somehow during an ICE restart
peter: The change should happen when the answer is applied
<mt_____> For DTLS, we agreed that we would create a new object when a new DTLS association is created (a=tls-id change, for instance)
Agreement is to go with the proposed resolution
<mt_____> do we have stats to put on these things?
Resolution is to *not* add getStats() to DTLS and ICE transport objects
<mt_____> I don't know about this, but can your role change?
<burn> +1 Randal. Id was only added to enable out of band negotiation
Resolution is to ignore the id value when in-band negotiation is used
Taylor will create a PR
Taylor proposes that an arbritary priority value should be mapped to the next valid value
bernard: Taylor will create the PR
Resolution is that the dtmf attribute is always set for senders of type audio, but insertDTMF method can fail
RESOLUTION: MST ids are not unique
<jib> Btw, there's the note on remote id's not being unique: http://w3c.github.io/webrtc-pc/#id
RESOLUTION: Is for Bernard to create a PR based on Tim's suggestion
<mt_____> A sequence of dictionaries that are "recommended" algorithms
Feature is implemented in firefox
<burn> Feature at risk just means you don't think you'll have at least two implementations
Resolution is that "Rollback" feature is *not* at risk
i.e. close issue 1215
bernard: Currently we don't have error handlers on senders and receivers
RESOLUTION: It's a best effort system