Seems like the reply route is a property of the transaction, and not of each respective branch. I wouldn't call it "bug" at the moment, but would call it a "limitation".
I second Juha's command - as soon as a branch is canceled I do not care about any following responses and it may be useful to drop them (to not interact with some logic waiting for responses on the new branch) - but different users have different scenarios ;-)
regards Klaus
On 11.10.2012 14:26, Alex Hermann wrote:
Hello,
i noticed thet in serial forking, replies to earlier branches arriving after sending out a new branch use the onreply_route of the later branch instead of the onreply_route set before sending the earlier branch....
How to reproduce:
set onreply_route to A
relay 1st branch
1st branch times out, internal 408 is created
tm send CANCEL to 1st branch
in failure route, onreply_route is set to B
relay 2nd branch
1st branch responds with 487, and goes into reply_route B instead of A
I think each branch should take the reply_route which was set before it got relayed and not pick up later changes meant for other branches.
How would i fix this?