You want to avoid to copy&paste redundant code in your branch-route, right? So the following snipped should do the same as if you were able to execute branch_route[1] after branch_route[3]:

route { if(/* initial invite */) t_on_branch("3"); else t_on_branch("1");

// ... }

// your branch-route helper route[20] { // some code executed by both branch routes }

// broute for common requests branch_route[1] { route(20); }

// broute for initial invites branch_route[3] { // some branch-specific stuff here, then:

route(20); }

Iñaki Baz Castillo wrote:

El Wednesday 17 October 2007 13:28:55 Andreas Granig escribió:

...

Hi,

You may also create a route[] for the common code and call that route from your branch routes.

I thougt that but have a doubt: if I call a "route(1)" from "branch_route[3]", what is processed in "route[1]"? all the branch together of each one?

So, if I do:

route { ... t_on_branch("1"); t_relay; }

branch_route[1] { if (is_method("INVITE") ) route(3); ... ... }

route[3] { ... operate_in_separate_branch here ¿?¿?¿? o in all branches together ?¿?¿?¿ }

What is processed in route[3] ?

Thanks.

If a "route[X]" is called from a "branch_route[Y]" then in that "route[X]" is just processed each separate branch and not all together as in main route, is it?

Yes, changes made in route[X] are only visible to the current branch being processed. Right, Daniel?

Andreas

El Wednesday 17 October 2007 14:48:36 Daniel-Constantin Mierla escribió:

Hello,

On 10/17/07 15:41, Andreas Granig wrote:

...

...

If a "route[X]" is called from a "branch_route[Y]" then in that "route[X]" is just processed each separate branch and not all together as in main route, is it?

Yes, changes made in route[X] are only visible to the current branch being processed. Right, Daniel?

yes. For example, this is a way to do rtp relaying only for natted branches, or to drop a branch if does not meet some conditions.

Ok, is nice to know it. I'll add it to the wiki since is nothing about it.

Thanks a lot.