ok, I see the difference, I will test this case as well before
releasing 3.1.5 and come back later with another message with more
details.
Cheers,
Daniel
On 9/14/11 3:17 PM, Vitaliy Aleksandrov wrote:
Hello. I found the difference between my config and
yours. I have t_newtran at the beginning of the request route for
all INVITE messages without to-tag.
Looks like without t_newtran all works fine.
As i understand t_relay creates a new transaction and while
first invite waiting for a db / radius or "100 Trying"
re-transmit may come. And till first invite call t_relay
re-transmitted copy will go through all request-route.
Could you pls explain why i can't call t_newtran() before
uac_req_send() ?
On Wed, Sep 14, 2011 at 12:31 AM,
Daniel-Constantin Mierla <miconda@gmail.com>
wrote:
Hello,
I did a small test myself with following config and avp(x)
was printed ok in failure route. All looks fine, so check
to see if you don't delete them directly somehow,
avp_print() should help printing all avps, as I said in
previous email.
On 9/13/11 5:02 PM, Daniel-Constantin Mierla wrote:
Hello,
On 9/13/11 3:04 PM, Vitaliy Aleksandrov wrote:
Hello,
i tried with a latest 3.1 and now avps works
after uac_req_send, but i can't get avp values
from onreply and failure routes.
do you mean no avp is available in onreply and
failure routes but they are available in request
route block up to t_relay()? Have you used
avp_print() from avpops to print the list of all
avps (at INFO level)?
Cheers,
Daniel
On Mon, Sep 12, 2011 at
10:21 AM, Daniel-Constantin Mierla <miconda@gmail.com>
wrote:
Hello,
the bug was associated with an internal tm
function used to send the request. Can you
try again with latest branch 3.1, over the
weekend I backported the fix. Let me know
if works now for you.
Thanks,
Daniel
On 9/9/11 3:33 PM, Vitaliy Aleksandrov
wrote:
There is a bug in UAC module.
uac_req_send doesn't restore avps.
I tried to backport your patch
to 3.1 but with no success.
Could you add this patch to a
new stable release ?