Hi Ovidiu,
I found the bug - it is a very spuky one with no obvious fix - by having
the t_relay() to work on an already created transaction, the
retransmission detection was bit (more) screwed up.....:(
regards,
bogdan
Ovidiu Sas wrote:
Hi Bogdan,
The problem is triggered when from mangling is involved and t_relay is
used in manual mode and 100 Trying is not sent.
An INVITE comes in. It's From headers is mangled and then is sent out
(using usrloc lookup). The INVITE is relayed using t_relay("0x01")
and no "100 Trying" is sent out.
On the first INVITE retransmission, the bug will occur.
Regards,
Ovidiu Sas
On 2/13/07, Ovidiu Sas <[EMAIL PROTECTED]> wrote:
Hi Bogdan,
Yes, the second INVITE is a retransmission.
The script is pretty simple: lcr followed by t_relay("0x01").
I just don't send back a 100 Trying. Could this cause issues?
Regards,
Ovidiu Sas
On 2/13/07, Bogdan-Andrei Iancu <[EMAIL PROTECTED]> wrote:
> Hi Ovidiu,
>
> second INVITE looks to be a retransmission. Are you using
t_newtran() +
> t_relay() ?? what is the script sequence for relaying?
>
> regards,
> bogdan
>
> Ovidiu Sas wrote:
> > Hi,
> >
> >
> > I got this bizarre 500 error (openser 1.2.0-pre4-tls (arm/linux)).
> >
> > 192.168.2.102 - CISCO IP PHONE
> > 192.168.2.22 - openser
> > 192.168.2.20 - asterisk
> >
> > Cisco --> openser --> asterisk
> >
> >
> > Here's the the log:
> > ERROR:tm:t_forward_nonack: no branch for forwarding
> > ERROR:tm:w_t_relay: t_forward_nonack failed
> > ERR: t_relay() in route[1]
> >
> >
> > and here's the ngrep:
_______________________________________________
Devel mailing list
[email protected]
http://openser.org/cgi-bin/mailman/listinfo/devel
