Fellows,
I started using the cvs version of ser instead of 0.8.10. But I'm facing some doubts on strict routing vs. loose routing.
In 0.8.10 I was using:
addRecordRoute();
rewriteFromRoute();
Now in CVS version, by default strict routing is disabled, right?
So, according to the rr module, I need to use record_route_strict, so I compiled rr module with it. Now, I'm using
record_route_strict() instead of addRecordRoute() and loose_route() instead of rewriteFromRoute();
Is this correct, or am I missing something?
Thanks in advance,
Guilherme.
Hello,
is there any reason why you need strict routing ? You should use loose routing instead (using record_route() function), strict routing has been deprecated. If you really need strict routing that record_route_strict is the right function.
Jan.
On 05-05 15:46, Gulherme Dal Pizzol wrote:
Fellows,
I started using the cvs version of ser instead of 0.8.10. But I'm facing some doubts on strict routing vs. loose routing.
In 0.8.10 I was using:
addRecordRoute();
rewriteFromRoute();
Now in CVS version, by default strict routing is disabled, right?
So, according to the rr module, I need to use record_route_strict, so I compiled rr module with it. Now, I'm using
record_route_strict() instead of addRecordRoute() and loose_route() instead of rewriteFromRoute();
Is this correct, or am I missing something?
Thanks in advance,
Guilherme.
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Jan,
If I'm working with RFC2543 compliant phones (such as ATA) may I use loose routing? Will cisco ata understand lr parameter and generate an ACK/BYE message correctly?
thanks again,
Guilherme.
Hello,
is there any reason why you need strict routing ? You should use loose routing instead (using record_route() function), strict routing has been deprecated. If you really need strict routing that record_route_strict is the right function.
Jan.
On 05-05 15:46, Gulherme Dal Pizzol wrote:
Fellows,
I started using the cvs version of ser instead of 0.8.10. But I'm facing some doubts on strict routing vs. loose routing.
In 0.8.10 I was using:
addRecordRoute();
rewriteFromRoute();
Now in CVS version, by default strict routing is disabled, right?
So, according to the rr module, I need to use record_route_strict, so I compiled rr module with it. Now, I'm using
record_route_strict() instead of addRecordRoute() and loose_route() instead of rewriteFromRoute();
Is this correct, or am I missing something?
Thanks in advance,
Guilherme.
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
On 05-05 16:27, Gulherme Dal Pizzol wrote:
Jan,
If I'm working with RFC2543 compliant phones (such as ATA) may I use loose routing? Will cisco ata understand lr parameter and generate an ACK/BYE message correctly?
Loose routing is backwards compatible, that means RFC2543 compliant devices should work with RFC3261 compliant servers.
But there are some user agents that strip the lr parameter when generating Route header fields (i.e. Record-Route contained lr parameter but Route created by the user-agent didn't). If that is your case then you can use modparam("rr", "enable_full_route", 1) which will insert ;lr=on instead of just ;lr which worked for us.
You should give loose routig a try and if it doesn't work for any reason then switch back to strict routing (please report such a misbehavior to me).
Jan.
Hello,
is there any reason why you need strict routing ? You should use loose routing instead (using record_route() function), strict routing has been deprecated. If you really need strict routing that record_route_strict is the right function.
Jan.
On 05-05 15:46, Gulherme Dal Pizzol wrote:
Fellows,
I started using the cvs version of ser instead of 0.8.10. But I'm facing some doubts on strict routing vs. loose routing.
In 0.8.10 I was using:
addRecordRoute();
rewriteFromRoute();
Now in CVS version, by default strict routing is disabled, right?
So, according to the rr module, I need to use record_route_strict, so I compiled rr module with it. Now, I'm using
record_route_strict() instead of addRecordRoute() and loose_route() instead of rewriteFromRoute();
Is this correct, or am I missing something?
Thanks in advance,
Guilherme.
Serusers mailing list serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers
Jan Janak writes:
But there are some user agents that strip the lr parameter when generating Route header fields (i.e. Record-Route contained lr parameter but Route created by the user-agent didn't). If that is your case then you can use modparam("rr", "enable_full_route", 1) which will insert ;lr=on instead of just ;lr which worked for us.
lr=on may help in case of broken UAs, but are you sure that all RFC3261 compliant AUs will understand lr=on as the lr parameter?
-- juha
On 06-05 08:01, Juha Heinanen wrote:
Jan Janak writes:
But there are some user agents that strip the lr parameter when generating Route header fields (i.e. Record-Route contained lr parameter but Route created by the user-agent didn't). If that is your case then you can use modparam("rr", "enable_full_route", 1) which will insert ;lr=on instead of just ;lr which worked for us.
lr=on may help in case of broken UAs, but are you sure that all RFC3261 compliant AUs will understand lr=on as the lr parameter?
No, I am not sure. I saw user agents doing very strange stuff so I am pretty sure that there will be at least one which will not handle this correctly.
The spec says that UAs should look if there is an lr parameter in URI and, IMHO, ;lr=on counts too because it is lr parameter, it just doesn't have empty body (but why would UAs look into the body when they just need to know if there is such parameter or not ?).
Anyway, if you find a UA that doesn't work with lr=on, please report it to me.
Jan.
Jan Janak writes:
The spec says that UAs should look if there is an lr parameter in URI and, IMHO, ;lr=on counts too because it is lr parameter, it just doesn't have empty body (but why would UAs look into the body when they just need to know if there is such parameter or not ?).
ok, thanks for the clarification. i'll check if kphone handles this correctly.
-- juha