- From: Robin Raymond <robin@hookflash.com>
- Date: Sun, 13 Apr 2014 16:09:24 -0400
- To: Emil Ivov <emcho@jitsi.org>
- CC: public-ortc@w3.org
- Message-ID: <534AEEF4.10105@hookflash.com>
That's what I've been talking about throughout that the local candidates are not changeable. However, there's a side effect. The RTP/RTCP often are required to share the same ufrag:password locally. The only way to get that to happen is to copy the credentials from an existing RtcIceListener object into a new RtcIceListener object since the web developer isn't allowed to set them directly. Now maybe you understand the context of where I was coming from... -Robin > Emil Ivov <mailto:emcho@jitsi.org> > April 13, 2014 at 3:56 PM > > Hey Robin, > > On 13 Apr 2014 9:48 PM, "Robin Raymond" <robin@hookflash.com > <mailto:robin@hookflash.com>> wrote: > > > > > > Hi Emil, > > > > I think I understand the fundamental confusion point. > > I think I do too. We might have been referring to different sets of > credentials. > > > ICE is used for consent and consent freshness. [BA] confirmed that it > > was not changed to Dtls. I think we both agree on this point. > > Yes. > > > However, it's my understanding that the browser generates the local > > ufrag:password and it's NOT changeable once created by the browser > > Yes, of course. The local aren't. In order for ICE to work however and > for ICE restarts to be possible, an application has to be able to set > new *remote* credentials. > > There is absolutely no need for an application (be it web or rich) to > exert control on the local credentials, other than requesting new ones. > > Emil > > --sent from my mobile >
Received on Sunday, 13 April 2014 20:09:54 UTC