Hello Benny, thank you very much for the insight! Does PJSIP provide some sort of probing mechanism (on the pjsua_level) to learn about the contact_rewrite_method supported by the particular server? If we are dealing with a countable number of well-known server manufacturers, one could install a small dictionary to look-up the manufactor-specific method on the application-level (or maybe even on the pjsua_level). - I suppose, one could also implement some kind of "heuristic" trial-error mechanism by e.g. 1) registering an account with PJSUA_CONTACT_REWRITE_METHOD, 2) automatically issuing a call to that account and interpret the server's reaction ...? (From my perspective it seems that such a facility would fit neatly on the "pjsua_" abstraction level... ;-) ?) What is the recommended tactic to deal with this situation? Again, thank you very much in advance! -Thomas On Dec 13, 2010, at 09:45 , Benny Prijono wrote: > On Sat, Dec 11, 2010 at 10:48 PM, Thomas Martin <tmemail at gmx.de> wrote: >> >> On the other hand: what is the real disadvantage of legacy (1)? >> > > Some servers from a well known equipment manufacturer will close the > TCP connection as soon as it sends 200/OK to unregister request, hence > method 1 doesn't work. Method 2 was created exactly because of this. > > On the other hand, another server from yet another well known VoIP > company don't have a clue that we're updating our contact using method > 2, and it thinks that we're doing registration with multiple Contacts, > and it will reject the registration since it doesn't do forking. And > of course, yet another server thinks the same, as you experienced. > > So each has a drawback. > > -Benny > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip at lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
