I am wondering how to expose and test the value of the expires parameter in a REGISTER request.
I am experimenting with openser as the basis for a home phone network. I use multiple devices with the same user ID. They register locally ( with no reply ) and with an external service provider. The contacts are mangled to show the public address of openser. Multiple registrations result in a single AOR at the external registrar. Incoming calls from the outside are forked and ring the local phones. Local phones can also call each other without the hairpin problem associated with STUN enabled phones.
The problem is that a softphone will deregister when it is closed or its profile changes. This would deregister the AOR at the external registrar. The remaining phones could not receive calls from the outside until they refreshed their registrations.
I would like to prevent deregistration at the external registrar unless the phone that was deregistering was the only remaining one. The first step would be to identify REGISTER messages where the expires value is equal to zero.
Any thoughts? Thanks Rob
Robert Dyck wrote:
I am wondering how to expose and test the value of the expires parameter in a REGISTER request.
I am experimenting with openser as the basis for a home phone network. I use multiple devices with the same user ID. They register locally ( with no reply ) and with an external service provider. The contacts are mangled to show the public address of openser. Multiple registrations result in a single AOR at the external registrar. Incoming calls from the outside are forked and ring the local phones. Local phones can also call each other without the hairpin problem associated with STUN enabled phones.
The problem is that a softphone will deregister when it is closed or its profile changes. This would deregister the AOR at the external registrar. The remaining phones could not receive calls from the outside until they refreshed their registrations.
I would like to prevent deregistration at the external registrar unless the phone that was deregistering was the only remaining one. The first step would be to identify REGISTER messages where the expires value is equal to zero.
Both 'Expires' header and 'expires' contact uri parameter have to be checked like e.g.
if ((is_present_hf("Expires") && $(hdr(Expires){s.int}) == 0) || ($(ct{param.value,expires}) == '0')) { # someone tries to unregister }
Have a look at http://www.openser.org/dokuwiki/doku.php/transformations:1.2.x if you're not familiar with the PV transformations introduced with 1.2.
/Christian
Any thoughts? Thanks Rob
Users mailing list Users@lists.openser.org http://lists.openser.org/cgi-bin/mailman/listinfo/users
Christian Schlatter wrote:
Robert Dyck wrote:
I am wondering how to expose and test the value of the expires parameter in a REGISTER request.
I am experimenting with openser as the basis for a home phone network. I use multiple devices with the same user ID. They register locally ( with no reply ) and with an external service provider. The contacts are mangled to show the public address of openser. Multiple registrations result in a single AOR at the external registrar. Incoming calls from the outside are forked and ring the local phones. Local phones can also call each other without the hairpin problem associated with STUN enabled phones.
The problem is that a softphone will deregister when it is closed or its profile changes. This would deregister the AOR at the external registrar. The remaining phones could not receive calls from the outside until they refreshed their registrations.
I would like to prevent deregistration at the external registrar unless the phone that was deregistering was the only remaining one. The first step would be to identify REGISTER messages where the expires value is equal to zero.
Both 'Expires' header and 'expires' contact uri parameter have to be checked like e.g.
if ((is_present_hf("Expires") && $(hdr(Expires){s.int}) == 0) || ($(ct{param.value,expires}) == '0')) { # someone tries to unregister }
Of course this only makes sense for REGISTER messages:
if (is_method("REGISTER")) ...
/Christian
Have a look at http://www.openser.org/dokuwiki/doku.php/transformations:1.2.x if you're not familiar with the PV transformations introduced with 1.2.
/Christian
Any thoughts? Thanks Rob
Users mailing list Users@lists.openser.org http://lists.openser.org/cgi-bin/mailman/listinfo/users
Users mailing list Users@lists.openser.org http://lists.openser.org/cgi-bin/mailman/listinfo/users
On Wednesday 31 October 2007, Christian Schlatter wrote:
Robert Dyck wrote:
I am wondering how to expose and test the value of the expires parameter in a REGISTER request.
I am experimenting with openser as the basis for a home phone network. I use multiple devices with the same user ID. They register locally ( with no reply ) and with an external service provider. The contacts are mangled to show the public address of openser. Multiple registrations result in a single AOR at the external registrar. Incoming calls from the outside are forked and ring the local phones. Local phones can also call each other without the hairpin problem associated with STUN enabled phones.
The problem is that a softphone will deregister when it is closed or its profile changes. This would deregister the AOR at the external registrar. The remaining phones could not receive calls from the outside until they refreshed their registrations.
I would like to prevent deregistration at the external registrar unless the phone that was deregistering was the only remaining one. The first step would be to identify REGISTER messages where the expires value is equal to zero.
Both 'Expires' header and 'expires' contact uri parameter have to be checked like e.g.
if ((is_present_hf("Expires") && $(hdr(Expires){s.int}) == 0) || ($(ct{param.value,expires}) == '0')) { # someone tries to unregister }
Have a look at http://www.openser.org/dokuwiki/doku.php/transformations:1.2.x if you're not familiar with the PV transformations introduced with 1.2.
I am indeed unfamiliar with PV transformations. I will have a look it. I was afraid I might have to do something ugly with regular expressions. I probably should not put off upgrading any longer.
Thanks, Rob
-
Christian Schlatter -
Robert Dyck