[Devel] presence and to uri

[Anca-Maria Vamanu]

Hello,

Thank you for reporting. I have fixed it in the last commit.

regards,

Anca Vamanu

Juha Heinanen wrote:

>i sent yesterday privately an email to anca about this, but thought that
>perhaps a broader discussion is appropriate, because it looks to me that
>there is a fundamental flaw in current presence implementation.
>
>while testing presence, i noticed that when processing subscribe
>requests, presence module seems to take presentity uri from to uri and
>not from the request uri.  for example, when twinkle sends subscribe
>
>SUBSCRIBE sip:jh at vm.test.fi SIP/2.0
>To: "Juha Heinanen" <sip:+358442345670 at test.fi>
>
>openser does not find presentity with username jh/domain vm.test.fi
>although it does exists.
>
>also, if state of presentity changes, twinkle does not get notification
>of it, i guess, because in active_watchers table twinkle is listed with
>to_user +358442345670/to_domain test.fi and there is no field in
>active_watchers table that would include presentity uri
>sip:jh at vm.test.fi that twinkle subscribed to.
>
>RFC 3265 is clear on how target of subscribe is determined:
>
>3.1.2. Identification of Subscribed Events and Event Classes
>
>   Identification of events is provided by three pieces of information:
>   Request URI, Event Type, and (optionally) message body.
>
>it makes never (except in register request) sense by presence server or
>proxy to examine to header, because to header has only local meaning.
>for example, my proxy may accept subscribe request where r-uri and to
>uri are both sip:my_mailbox and then rewrite request uri to
>sip:jh at vm.test.fi.
>
>another (minor) thing that i noticed while reading presence source code
>is that it uses parse_uri functions to parse to header.  these should
>be changed to more efficient parse_to_uri, which checks if the uri has
>already been parsed.
>
>-- juha
>
>_______________________________________________
>Devel mailing list
>Devel at openser.org
>http://openser.org/cgi-bin/mailman/listinfo/devel
>
>  
>