Author: rco
Date: 2008-04-18 14:02:52 +0200 (Fri, 18 Apr 2008)
New Revision: 896
Modified:
trunk/core/plug-in/sipctrl/trans_layer.cpp
Log:
- fixed bug incorrectly fixed in r895.
Modified: trunk/core/plug-in/sipctrl/trans_layer.cpp
===================================================================
--- trunk/core/plug-in/sipctrl/trans_layer.cpp 2008-04-17 13:19:56 UTC (rev
895)
+++ trunk/core/plug-in/sipctrl/trans_layer.cpp 2008-04-18 12:02:52 UTC (rev
896)
@@ -893,8 +893,19 @@
case TS_CALLING:
case TS_PROCEEDING:
+ // TODO:
+ // we should take care of 200 ACKs
+ // - on first reply:
+ // - save to-tag.
+ // - use route-set included in the INV req (if applicable).
+ // - save ACK for retransmitions.
+ // - compare to-tag on subsequent replies.
+ // - (if different):
+ // - (generate new 200 ACK based on reply).
+ // - (send BYE (check for existing UAC trans)).
+ // - else:
+ // - re-transmit ACK.
t->state = TS_TERMINATED;
- // TODO: should we assume 200 ACK retransmition?
bucket->remove_trans(t);
goto pass_reply;
@@ -954,12 +965,10 @@
t->state = TS_COMPLETED;
if(t->msg->u.request->method == sip_request::INVITE){
- //TODO: set G timer ?
t->reset_timer(STIMER_G,G_TIMER,bucket->get_id());
t->reset_timer(STIMER_H,H_TIMER,bucket->get_id());
}
else {
- //TODO: set J timer ?
// 64*T1_TIMER if UDP / 0 if !UDP
t->reset_timer(STIMER_J,J_TIMER,bucket->get_id());
}
@@ -976,7 +985,7 @@
//
// In this stack, the transaction layer
// takes care of re-transmiting the 200 reply
- // in a UAS INVITE transaction. The core above
+ // in a UAS INVITE transaction. The code above
// is commented out and shows the behavior as
// required by the RFC.
//
@@ -987,8 +996,7 @@
}
else {
t->state = TS_COMPLETED;
- // TODO: set J timer
- // 0 if !UDP
+ // Only for unreliable transports.
t->reset_timer(STIMER_J,J_TIMER,bucket->get_id());
}
}
@@ -1170,7 +1178,7 @@
void trans_layer::retransmit(sip_trans* t)
{
assert(transport);
- if(!t->retr_buf || t->retr_len){
+ if(!t->retr_buf || !t->retr_len){
// there is nothing to re-transmit yet!!!
return;
}
_______________________________________________
Semsdev mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/semsdev
