Fw: Re: [Serusers] exec_msg and exec_uri integration

Cerrajetto cerrajetto at pyme.net
Fri Jul 18 10:38:23 CEST 2003


Hello!, 

OK, it's cool, this way we can obtain all headers and then "decide" where to 
rewrite the URI.

We would like to do some statistic control: our external program "decides" 
where to route, and writes in a Mysql DB some fields: from, to, date/time, 
interesting extra headers, ...

Is there a way to know date/time of hangup?. When a call hangs up, it 
generates sip info: is it possible to detect this info in ser.cfg to launch 
an external script (via exec_msg) ?. If possible, this script can update a 
table in DB with hangup information.

Thank you,
Cerrajetto.

---------- Forwarded Message -----------
From: Jiri Kuthan <jiri at iptel.org>
To: "Cerrajetto" <cerrajetto at pyme.net>, serusers at lists.iptel.org
Sent: Tue, 15 Jul 2003 09:48:14 +0200
Subject: Re: [Serusers] exec_msg and exec_uri integration

At 06:26 PM 7/14/2003, Cerrajetto wrote:
>Hello,
>
>We are testing exec_msg and exec_uri.
>
>exec_msg sends via standard input all sip headers and contents, but exec_uri 
>only sends current uri ... 

Almost all parts of SIP message are passed in environment variables,
see our doc: http://www.iptel.org/ser/doc/prerelease/c1088.html#USINGEXEC,
the last figure in example 2 (printenv).

>Therefore, we can only write external scripts to 
>redirect on a "current_uri" basis. We would like to integrate "exec_msg" 
>and "exec_uri" features, is it possible?

As said before, one alternative would be to serve yourself from env-vars.

The other way is writing a module which merges both functionalities. That
would take slightly more work than exec as is now -- exec relies on 'popen',
which makes lot of things for you in a portable way. It unfortunately also
sets limitations which made me split the module in two separate actions.

>We've found that after a change in ser.cfg it is needed to restart SER to 
>apply changes, is there a way to force a "reload" of ser.cfg?

There is not config-reload option. Not really too bad as user-loc database
is kept persistently, transactions can complete statelessly and that's all
state you have. Nevertheless, it is definitely a nice convenience feature
which should show up in SER some day.

-Jiri
------- End of Forwarded Message -------



More information about the sr-users mailing list