because i didn't think of it?

:-)

cheers, and thanx

Bogdan-Andrei Iancu wrote:

Kanakatti Mahesh Subramanya wrote:

...

I totally agree.

Personally, I've been ripping out large chunks of my config file, and replacing it with varioud avpops based stuff - it may be *slightly* more convoluted, but it is completely database configurable, and hence easier to maintain

re: rr, its actually neither. Its a static, which i'm using to identify NAT'ed clients to account for re-INVITEs

why don't you use the append_rr_param() function to added the NAT identifier to the RR hdr? do a normal record_route() and if nat is detected, add the param

regards, bogdan

...

cheers

Bogdan-Andrei Iancu wrote:

...

Hi Kanakatti,

I'm for m4'ing (for the moment) for two reasons:

1. since we can do it with m4, I prefer to use the cycles for

something still missing; 2) using dynamic parameters (like via AVPs) reduce the performance - time to look for the AVP and most critical the impossibility of calling fixup functions at startup (which will force to fix params at runtime, for each processing).

regarding the record_route_preset() - do you use dynamic parameters or dynamic rr URI?

regards, bogdan