26 Sep
2006
26 Sep
'06
7:30 a.m.
Hi!
loose_route is true if the received request contains routes - either
loose routed or strict routed. Thus, the naming is wrong, but this has
historical reasons. The loose route function also removes the Route
headers which points to the proxy itself (the matching is done via the
IP addresses the proxy is listening on, and via the configured 'alias')
and performs translation from loose-routing to strict-routing and
vice-versa if the needed.
If the message does not contain a Route set, loose_route return false.
In usual cases, this means that for out-of-dialog requests the
loose_route function returns false whereas for in-dialog requests (which
usually have route headers due to record-route) the loose_route function
returns true.
regards
klaus
Jeremie Le Hen wrote:
Hi list,
in numerous example configuration files, I saw the following statement
(where "route(1)" is the final routing block):
% if (loose_route()) {
% route(1);
% }
%
% ...
%
% route(1);
It works for me, but I am just wondering what causes the loose_route()
function to return a failure ? In this case, it seems number of actions
used to be performed before calling the final message routing block.
Naturally, I checked the rr module documentation, but the descriptions is
very small and only refers to RFC 3161. I have read some parts of this
RFC, thus I know what loose routing means (basically, it uses always the
same URI all along the proxy string, removing IP addresses in the Route
header gradually, contrariwise to strict routing which puts the real URI
at the end of the Route header and use the next proxy as URI). In spite
of this, I cannot figure why this function should fail.
I'm pretty confident this is to help for the scripting logic, and I was
supplied with some explanations.
Thank you.
Regards,