On 01/12/14 10:45, Olle E. Johansson wrote: > On 01 Dec 2014, at 10:42, Daniel-Constantin Mierla <mico...@gmail.com> wrote: > >> On 01/12/14 02:23, Juha Heinanen wrote: >>> Juha Heinanen writes: >>> >>>> it might be possible to store in branch route request uris to an avp and >>>> then select the correct one based on the last field of top via branch, >>>> but that is a far too complicated solution for such a simple task as >>>> finding out the request uri of replied request. therefore, i must be >>>> missing something. >>> another one would be to store $ru in branch route to htable using >>> transaction and branch index as key and then get it from there in onreply >>> route. still too complicated. >> I think indeed that would be an option. Looks like nobody needed this >> value so far ... >>> would it be possible to define $T(request_uri)? >> I think the best way is to extend $T_branch(...) class of pvs to have >> $T_branch(ruri), or maybe even $T_branch($pv) and have $T_branch($ru) >> for what you need. > Do you mean the "winning" branch? > > If we end up in failure route, there can be multiple branches ending up > with the same answer, like two phones sending 486. How do we handle > finding the ruri's of those? For a parallel forking, indeed, there can be such case with same code in many branches. We have to look at tm code, I guess Jiri Kuthan or Andrei Pelinescu implemented it long time ago (as I don't remember any work recently on that area) and perhaps in this case it selects the first or the last with the same code ... Cheers, Daniel
-- Daniel-Constantin Mierla http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda _______________________________________________ SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list sr-users@lists.sip-router.org http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
