- From: Bernard Aboba <Bernard.Aboba@microsoft.com>
- Date: Thu, 4 Dec 2014 21:49:07 +0000
- To: Peter Thatcher <pthatcher@google.com>
- CC: "public-ortc@w3.org" <public-ortc@w3.org>
Received on Thursday, 4 December 2014 21:49:37 UTC
Peter Thatcher said: Also, it makes me wonder: what happens to the IceTransports that are tied to the IceGatherer? What are they supposed to do? Do they all go back to the "connecting" state and out of the "completed" state? Would one expect to change one object and then the state of a bunch of other objects change like that? [BA] The current text does not make it clear whether .start can be called again with an existing IceGatherer and if so, what the expected behavior is. If .start is called again with an existing IceGatherer, the ufrag/password remains unchanged, so that using the term “restart” in the [RFC5245] sense is probably not appropriate. Moreover, depending on the state of the IceGatherer, some of the originally gathered ICE candidates may have been pruned; if so, what happens when calling .start again? Also, it is not clear what the expected behavior is if a connectivity check arrives prior to calling .start initially. It seems logical for the IceGatherer to respond to the checks, but this is not stated.
Received on Thursday, 4 December 2014 21:49:37 UTC