Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.
We've got a situation where we use t_send_reply() to generate early 100 Trying, before an operation that might have a small delay. The result of this operation might lead to a change which adds parameters to ruri-user. This works as expected for normal operations, but when calling ds_next_dst()/ds_next_domain() the change to $rU is lost.
I then noticed that this is because t_send_reply() indirectly does a t_newtran() before change to $rU. If it's done in the opposite order, everything works as expected.
A workaround is to use for example sl_send_reply() and t_set_auto_inv_100(0), which works for me.
However, I'm unsure if this revert of user is expected behaviour?
$rU = $rU + ";npdi";
..
ds_select_dst(..);
then trigger a condition which executes failure_route and then use ds_next_dst() to fork serially.
$rU is then reverted to the original value in outgoing INVITE.
Reproduced in master.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.