On Tue, May 20, 2008 at 2:39 AM, Alan J. Bond <alan.bond at wintology.com> wrote: > Hi Benny, > > Sorry for late feedback. Have tested the new code against SER with fix_nated_contact() configuration option, with and without add_xuid_param set. Attached is log of registration with add_xuid_param = 1. I like the detection of SER's changing of the contact address and re-registration. > > As you'll see, SER returns our x-uid GUID as expected and the fallback matching works great. Interestingly SER also lops 5 seconds off the requested expiry, presumably to account for latency per our previous discussions. > > Will run a further test with multiple clients. Won't bore everyone with results if as expected it all works fine. > > A really excellent solution :-) > Thanks for the feedback, Alan. I've had various scenarios tested in our registration test vectors [1], including multiple bindings, modified and unmodified, and even scenario when no Contact is returned, so I'm pretty confident that it'll work as expected. Of course nothing beats real world tests so if you have time to spare for it then I'll appreciate your feedback. Cheers Benny [1] http://trac.pjsip.org/repos/browser/pjproject/trunk/pjsip/src/test-pjsip/regc_test.c#L862
