Sections 2.2 and 2.3: RTCDtlsTransport #38
A RTCDtlsTransport instance is optionally constructed from an RTCIceTransport object or an RTCIceTransport is automatically constructed.
[BA] Should this be "or an RTCDtlsTransport is automatically constructed"?
2.3 Interface Definition
[Constructor(optional RTCIceTransport transport)]
Assuming we have a DTLS transport constructed without an ice transport object as an argument, what happens when an app calls start() on the DTLS transport? Is an RTCIceTransport object really an optional argument to the constructor?
The text was updated successfully, but these errors were encountered:
I do think it's an optional argument. Start will have to do nothing until it's attached later to an ICE transport. It's kind of in an "awaiting attachment" state. Once attached, it should continue to perform it's operations. If reattached to another ICE transport it should continue its internal state exactly where it left off. If it becomes detached it should freeze its state "as is" and await attachment again to any ICE transport.
…c#27 Support for control of quality, resolution, framerate and layering added, as described inhttps://github.com/w3c/issues/31 RTCRtpListener object added and figure in Section 1 updated, as described in w3c#32 More complete support for RTP and Codec Parameters added, as described in w3c#33 Data Channel transport problem fixed, as described in w3c#34 Various NITs fixed, as described in w3c#37 Section 2.2 and 2.3 issues fixed, as described in w3c#38 Default values of some dictionary attributes added, to partially address the issue described in w3c#39 Support for ICE TCP added, as described in w3c#41 Fixed issue with sequences as attributes, as described in w3c#43 Fix for issues with onlocalcandidate, as described in w3c#44 Initial stab at a Stats API, as requested in w3c#46 Added support for ICE gather policy, as described in w3c#47