Hello,

On 08/21/06 11:42, Gerry wrote:

Good question. I find this restriction very annoying, too. Sometimes you can get around by checking the things in the forward route and using setflag() / resetflag() / isflagset() .

Greetings, Gerry.

Federico Giannici schrieb:

...

Is there a reason why the avp_check() function cannot be used in the ONREPLY_ROUTE, while it can be used for example in the FAILURE_ROUTE?

Using OpenSER 1.1.0 from CVS of a few days ago.

this restriction comes because the replies are not processed under mutex, therefore the access to avps is not synchronized. We look for a solution to this issue, since it affects other features we want to have.

Cheers, Daniel

...

Thanks.

---

Users mailing list Users@openser.org http://openser.org/cgi-bin/mailman/listinfo/users

Federico Giannici wrote:

Daniel-Constantin Mierla wrote:

...

Hello,

On 08/21/06 11:42, Gerry wrote:

...

Good question. I find this restriction very annoying, too. Sometimes you can get around by checking the things in the forward route and using setflag() / resetflag() / isflagset() .

Greetings, Gerry.

Federico Giannici schrieb:

...

Is there a reason why the avp_check() function cannot be used in the ONREPLY_ROUTE, while it can be used for example in the FAILURE_ROUTE?

Using OpenSER 1.1.0 from CVS of a few days ago.

this restriction comes because the replies are not processed under mutex, therefore the access to avps is not synchronized. We look for a solution to this issue, since it affects other features we want to have.

So, it shouldn't be usable in the FAILURE_ROUTEs too. Or not?

from failure route is safe to use since the route is protected by a lock from TM - a transaction has only one final reply, but have several provisional replies that can be processed in parallel.

regards, bogdan