Hi Julius, I was thinking about something like that, but most gateway or billing system vendors send/require Service-Type = Login-User. You can either write a small authenticator derived from RadAuth/RadAliasAuth that overrides OnSendPDU method and replaces Login-User with Voice or you can distinguish RRQ from ARQ by absence/presence of h323-conf-id attribute, for example (or h323-call-type). The first way you will get what you want and can benefit from further gk updates, because your derived class would be in a separate file. ----- Original Message ----- From: "Julius Stavaris" <TJST@bite.lt> Sent: Sunday, October 19, 2003 3:02 PM >Hi, > I have one suggestion regarding radius atribute Service-Type sent by RRQ's & ARQ's. > At the moment it is : > > For RRQ: "Login-User" > For ARQ: Service-Type - Login-User (for ARQs from originating endpoint) > or Call-Check (for ARQs from answering endpoint) > > From my point of view the problem is that we do not have single radius attribute to make the distinction between ARQ & >RRQ. The only difference is attributes present in request, but I think it would be easier to have single attribute whitch clearly >points to the meaning (ARQ or RRQ) of the request. > Lets say we could use: > > For RRQ: Service-Type - Login-User > For ARQ originate: Service-Type - Voice (Outbound-User would be another candidate) > For ARQ originate: Service-Type - Call-Check > > Julius ------------------------------------------------------- This SF.net email sponsored by: Enterprise Linux Forum Conference & Expo The Event For Linux Datacenter Solutions & Strategies in The Enterprise Linux in the Boardroom; in the Front Office; & in the Server Room http://www.enterpriselinuxforum.com _______________________________________________ List: Openh323gk-users@lists.sourceforge.net Archive: http://sourceforge.net/mailarchive/forum.php?forum_id=8549 Homepage: http://www.gnugk.org/
