17 Feb
2009
17 Feb
'09
9:23 p.m.
On 02/17/2009 07:53 PM, Jan Janak wrote:
Indeed, there will be some duplicated modules for a while, so in
specific cases/config, a set of modules will be from one side only. This
is fine for me with the first integration version.
Cheers,
Daniel
--
Daniel-Constantin Mierla
http://www.asipto.com
On 17-02 18:50, Jan Janak wrote:
> On 17-02 19:35, Daniel-Constantin Mierla wrote:
>
>> Hello Jan,
>>
>> lot of work, thanks.
>>
>> Practically now main components are integrated, with the reserve that
>> the statistics part from kamailio would need a better integration with
>> atomics ops in ser. It is why I haven't merge it in master branch yet.
>>
>> Still to do:
>> - integrate the other db drivers, as you mentioned
>> - core extensions in kamailio - Andrei already did the while and
>> completing switch is on the pipe
>> - tm extensions to get the other kamailio modules depending on tm work
>> properly
>>
>> Then we are done ... looks not much but still some work ...
>>
> I don't think that's it. You'll probably run into many more issues
definitively we are not there yet, otherwise we would be preparing a
release, what I mean was that even looks trivial at first sight, still
some work to do ...
> when you
> try to port some modules to sip-router. So far I hit:
>
> * Missing functions in core (hash functions from kamailio)
>
they are easy to fix. I did it with some defines in PV api and
statistics, but would be better to add some wrappers for full compatibility.
> * Incompatible avp data structures and
functions
>
This should be updated in Kamailio side to use the PV API whenever is
possible, so the AVP structure is hidden. There are cases when
operations require direct access (e.g., avopos module).
* ival
versus i in fparam/gparam
* Several modules in ser depend on domain modules which is different from
the domain module in kamailio.
* Several modules depend on xlog which is again different in kamailio.
I forgot to mention that the last two bullets are not show stoppers.