[Users] PIC/SER PA presence module port?

Bogdan-Andrei Iancu bogdan at voice-system.ro
Wed Nov 30 17:41:23 CET 2005

Hi Joachim,

for the moment there are no immediate plans to invest work in a present 
agent....and this mainly due the lack of standardisation in the area. 
each SIP client has it;s one interpretation of how presence should work 
and building a PA that fits only to some clients (even those subject to 
changes) makes no sense..


Joachim Fabini wrote:

>>>Did anyone succeed (or, at least, try) to port the new SER 
>>>pa module 
>>>(Jamey Hicks' additions for PIC/SER trial) to openSER?
>>the presence stuff will be approached in the future in 
>>OpenSER, too. I have seen a lot of changes and several modes 
>>of operating for the new PA in SER. 
>Hmmh, let's go a little bit into details: Are there plans to 
>port SER's pa module implementation or is there consensus to 
>implement something from the scratch? 
>I agree with you that after the PIC/SER trial there is a lot 
>of ongoing work concerning SER's pa module. What they still
>did not address is the B2BUA issue (imho, please correct me
>if I'm wrong) but they have many of the basic presence 
>features already working.
>>Not sure yet what sip 
>>client can be used for testing, I have seen remarks only 
>>about eyebeam.
>Eyebeam is afaik one of the more flexible ones, probably a 
>good choice. Supports both, peer-to-peer and server-based 
>presence, which should be sufficient to test most of the
>possible scenarios.
>Afaik the PIC/SER used it for their trial.
>>Several days ago SER changed the indexing style for user 
>>location records, using now unique ids, otherwise the 
>>interfaces are still the same.
>Just had a short look at the sources - it shows up that
>having the SER pa module included into OpenSER is not
>just a matter of replacing the pa module sources. The 
>SER pa developers moved some of the functionality 
>(like XCAP) into libraries positioned below the root 
>folder. The pa module depends on some of these library 
>headers. Unfortunately, even satisfying these 
>dependencies by copying the lib folder is not sufficient. 
>Beside some macros that can easily be copied or replaced, 
>the SER folks seem to have changed some core code (e.g 
>added new member sipifmatch to struct sip_msg). That's 
>developer list stuff, so I stopped here... ;) 
>My conclusion: it should be feasible to eliminate all 
>compiler errors an integrate the pa module with openser. 
>Before starting integration it's advisable to have core 
>programmer feedback on side effects of changing central 
>structs like sip_msg. And I do not see major sense in 
>doing this work before the new SER pa module is stable 
>Or, is someone already working on a new pa module 
>implementation for OpenSER?
>best regards
>Users mailing list
>Users at openser.org

More information about the Users mailing list