I just did an experiment, I delayed the addIceCandidate calls in the computer that makes the answer and that reversed success conditions.
This means that it's not dependent on who makes the answer or offer, but dependent on the order in which the ice candidates are added. On Fri, Nov 14, 2014 at 10:06 PM, Mario Carbajal <[email protected] > wrote: > Yes, it does. > > On Fri, Nov 14, 2014 at 9:58 PM, Eric Rescorla <[email protected]> wrote: > >> Can you confirm that it fails with two chromes and with two firefoxes? >> >> On Fri, Nov 14, 2014 at 2:45 PM, Mario Carbajal <[email protected]> >> wrote: >> >>> The signaling of my project being wrong was my first bet, but it seems to >>> happen everywhere. >>> http://webrtcdevelop.appspot.com/ and https://www.sharefest.me/ for >>> example. >>> >>> On Fri, Nov 14, 2014 at 7:35 PM, Eric Rescorla <[email protected]> wrote: >>> >>> > It shouldn't matter. I wonder if something is going wrong with your >>> > signaling >>> > system. Which calling site are you using? >>> > >>> > -Ekr >>> > >>> > >>> > On Fri, Nov 14, 2014 at 10:56 AM, <[email protected]> wrote: >>> > >>> >> I have two computers trying to connect to each other with webrtc >>> >> (datachannel and STUN only). >>> >> Computer A with no NAT or firewall. >>> >> Computer B behind a router with NAT. >>> >> When A creates an offer and B the answer the connection is successful. >>> >> However, when B creates an offer and A creates the answer the >>> connection >>> >> fails. >>> >> >>> >> Shouldn't the connection success/failure be independent of who makes >>> the >>> >> offer and who makes the answer? >>> >> >>> >> If it's not independent, in order to maximize the chances of two users >>> >> being able to connect to each other webrtc applications would need to >>> >> attempt both combinations... >>> >> >>> >> I've tested these two computers with many different webrtc apps, in >>> both >>> >> firefox and chrome, it's always the same result. >>> >> _______________________________________________ >>> >> dev-media mailing list >>> >> [email protected] >>> >> https://lists.mozilla.org/listinfo/dev-media >>> >> >>> > >>> > >>> _______________________________________________ >>> dev-media mailing list >>> [email protected] >>> https://lists.mozilla.org/listinfo/dev-media >>> >> >> > _______________________________________________ dev-media mailing list [email protected] https://lists.mozilla.org/listinfo/dev-media
