freeswitch/libs/sofia-sip/libsofia-sip-ua/tport/tport_type_tcp.c

558 lines
14 KiB
C
Raw Normal View History

/*
* This file is part of the Sofia-SIP package
*
* Copyright (C) 2006 Nokia Corporation.
*
* Contact: Pekka Pessi <pekka.pessi@nokia.com>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public License
* as published by the Free Software Foundation; either version 2.1 of
* the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
* 02110-1301 USA
*
*/
/**@CFILE tport_type_tcp.c TCP Transport
*
* See tport.docs for more detailed description of tport interface.
*
* @author Pekka Pessi <Pekka.Pessi@nokia.com>
* @author Martti Mela <Martti.Mela@nokia.com>
*
* @date Created: Fri Mar 24 08:45:49 EET 2006 ppessi
*/
#include "config.h"
#include "tport_internal.h"
#if HAVE_NETINET_TCP_H
#include <netinet/tcp.h>
#endif
#ifndef SOL_TCP
#define SOL_TCP IPPROTO_TCP
#endif
#include <stdlib.h>
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
#include <string.h>
#include <time.h>
#include <assert.h>
#include <errno.h>
#include <limits.h>
#if HAVE_FUNC
#elif HAVE_FUNCTION
#define __func__ __FUNCTION__
#else
static char const __func__[] = "tport_type_tcp";
#endif
/* ---------------------------------------------------------------------- */
/* TCP */
tport_vtable_t const tport_tcp_vtable =
{
/* vtp_name */ "tcp",
/* vtp_public */ tport_type_local,
/* vtp_pri_size */ sizeof (tport_primary_t),
/* vtp_init_primary */ tport_tcp_init_primary,
/* vtp_deinit_primary */ NULL,
/* vtp_wakeup_pri */ tport_accept,
/* vtp_connect */ NULL,
/* vtp_secondary_size */ sizeof (tport_t),
/* vtp_init_secondary */ tport_tcp_init_secondary,
/* vtp_deinit_secondary */ NULL,
/* vtp_shutdown */ NULL,
/* vtp_set_events */ NULL,
/* vtp_wakeup */ NULL,
/* vtp_recv */ tport_recv_stream,
/* vtp_send */ tport_send_stream,
/* vtp_deliver */ NULL,
/* vtp_prepare */ NULL,
/* vtp_keepalive */ NULL,
/* vtp_stun_response */ NULL,
/* vtp_next_secondary_timer*/ tport_tcp_next_timer,
/* vtp_secondary_timer */ tport_tcp_timer,
};
tport_vtable_t const tport_tcp_client_vtable =
{
/* vtp_name */ "tcp",
/* vtp_public */ tport_type_client,
/* vtp_pri_size */ sizeof (tport_primary_t),
/* vtp_init_primary */ tport_tcp_init_client,
/* vtp_deinit_primary */ NULL,
/* vtp_wakeup_pri */ NULL,
/* vtp_connect */ NULL,
/* vtp_secondary_size */ sizeof (tport_t),
/* vtp_init_secondary */ tport_tcp_init_secondary,
/* vtp_deinit_secondary */ NULL,
/* vtp_shutdown */ NULL,
/* vtp_set_events */ NULL,
/* vtp_wakeup */ NULL,
/* vtp_recv */ tport_recv_stream,
/* vtp_send */ tport_send_stream,
/* vtp_deliver */ NULL,
/* vtp_prepare */ NULL,
/* vtp_keepalive */ NULL,
/* vtp_stun_response */ NULL,
/* vtp_next_secondary_timer*/ tport_tcp_next_timer,
/* vtp_secondary_timer */ tport_tcp_timer,
};
static int tport_tcp_setsndbuf(int socket, int atleast);
int tport_tcp_init_primary(tport_primary_t *pri,
tp_name_t tpn[1],
su_addrinfo_t *ai,
tagi_t const *tags,
char const **return_culprit)
{
int socket;
socket = su_socket(ai->ai_family, ai->ai_socktype, ai->ai_protocol);
if (socket == INVALID_SOCKET)
return *return_culprit = "socket", -1;
tport_tcp_setsndbuf(socket, 64 * 1024);
return tport_stream_init_primary(pri, socket, tpn, ai, tags, return_culprit);
}
int tport_stream_init_primary(tport_primary_t *pri,
su_socket_t socket,
tp_name_t tpn[1],
su_addrinfo_t *ai,
tagi_t const *tags,
char const **return_culprit)
{
pri->pri_primary->tp_socket = socket;
/* Set IP TOS if set */
tport_set_tos(socket, ai, pri->pri_params->tpp_tos);
#if defined(__linux__)
/* Linux does not allow reusing TCP port while this one is open,
so we can safely call su_setreuseaddr() before bind(). */
su_setreuseaddr(socket, 1);
#endif
if (tport_bind_socket(socket, ai, return_culprit) == -1)
return -1;
if (listen(socket, pri->pri_params->tpp_qsize) == SOCKET_ERROR)
return *return_culprit = "listen", -1;
#if !defined(__linux__)
/* Allow reusing TCP sockets
*
* On Solaris & BSD, call setreuseaddr() after bind in order to avoid
* binding to a port owned by an existing server.
*/
su_setreuseaddr(socket, 1);
#endif
pri->pri_primary->tp_events = SU_WAIT_ACCEPT;
pri->pri_primary->tp_conn_orient = 1;
return 0;
}
int tport_tcp_init_client(tport_primary_t *pri,
tp_name_t tpn[1],
su_addrinfo_t *ai,
tagi_t const *tags,
char const **return_culprit)
{
pri->pri_primary->tp_conn_orient = 1;
return 0;
}
int tport_tcp_init_secondary(tport_t *self, int socket, int accepted,
char const **return_reason)
{
2013-04-25 18:33:17 +00:00
int val = 1;
self->tp_has_connection = 1;
self->tp_params->tpp_idle = UINT_MAX;
2013-04-25 18:33:17 +00:00
if (setsockopt(socket, SOL_TCP, TCP_NODELAY, (void *)&val, sizeof val) == -1)
return *return_reason = "TCP_NODELAY", -1;
2013-04-25 18:33:17 +00:00
#if defined(SO_KEEPALIVE)
setsockopt(socket, SOL_SOCKET, SO_KEEPALIVE, (void *)&val, sizeof val);
#endif
val = 30;
#if defined(TCP_KEEPIDLE)
setsockopt(socket, SOL_TCP, TCP_KEEPIDLE, (void *)&val, sizeof val);
#endif
#if defined(TCP_KEEPINTVL)
setsockopt(socket, SOL_TCP, TCP_KEEPINTVL, (void *)&val, sizeof val);
#endif
if (!accepted)
tport_tcp_setsndbuf(socket, 64 * 1024);
return 0;
}
static int tport_tcp_setsndbuf(int socket, int atleast)
{
#if SU_HAVE_WINSOCK2
/* Set send buffer size to something reasonable on windows */
int size = 0;
socklen_t sizelen = sizeof size;
if (getsockopt(socket, SOL_SOCKET, SO_SNDBUF, (void *)&size, &sizelen) < 0)
return -1;
if (sizelen != sizeof size)
return su_seterrno(EINVAL);
if (size >= atleast)
return 0; /* OK */
return setsockopt(socket, SOL_SOCKET, SO_SNDBUF,
(void *)&atleast, sizeof atleast);
#else
return 0;
#endif
}
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
/** Return span of whitespace from buffer */
static inline size_t ws_span(void *buffer, size_t len)
{
size_t i;
char const *b = buffer;
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
for (i = 0; i < len; i++) {
if (b[i] != '\r' && b[i] != '\n' && b[i] != ' ' && b[i] != '\t')
break;
}
return i;
}
/** Receive from stream.
*
* @retval -1 error
* @retval 0 end-of-stream
* @retval 1 normal receive
* @retval 2 incomplete recv, recv again
*
*/
int tport_recv_stream(tport_t *self)
{
msg_t *msg;
ssize_t n, N, veclen;
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
int err, initial;
msg_iovec_t iovec[msg_n_fragments] = {{ 0 }};
N = su_getmsgsize(self->tp_socket);
if (N == 0) {
if (self->tp_msg)
msg_recv_commit(self->tp_msg, 0, 1);
return 0; /* End of stream */
}
if (N == -1) {
err = su_errno();
SU_DEBUG_1(("%s(%p): su_getmsgsize(): %s (%d)\n", __func__, (void *)self,
su_strerror(err), err));
return -1;
}
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
initial = self->tp_msg == NULL;
memset(&self->tp_ptime, 0, sizeof self->tp_ptime);
while (initial && N <= 8) { /* Check for whitespace */
char crlf[9];
size_t i;
n = su_recv(self->tp_socket, crlf, N, MSG_PEEK);
i = ws_span(crlf, n);
if (i == 0)
break;
n = su_recv(self->tp_socket, crlf, i, 0);
if (n <= 0)
return (int)n;
SU_DEBUG_7(("%s(%p): received keepalive (total %u)\n", __func__,
(void *)self, self->tp_ping));
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
N -= n, self->tp_ping += n;
sync to current darcs tree: Tue Nov 20 11:46:34 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * nua_register.c: silenced warning with URL_INIT_AS() on Sun CC Tue Nov 20 11:47:06 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * test_register.c: fixed authentication caching issue Tue Nov 20 12:09:19 EST 2007 Pekka.Pessi@nokia.com * tport: tport_name_is_resolved() now uses host_is_ip_address() Tue Nov 20 12:53:54 EST 2007 Pekka.Pessi@nokia.com * tport.c: fixed EXPENSIVE unresolved bug. Wed Nov 21 07:10:40 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * test_tport.c: skipping tests on TLS if send fails. Wed Nov 21 11:46:42 EST 2007 Pekka.Pessi@nokia.com * test_nta.c: removed dead code Thu Nov 22 08:42:14 EST 2007 Pekka.Pessi@nokia.com * test_nta.c: refactored client side tests Thu Nov 22 09:39:45 EST 2007 Pekka.Pessi@nokia.com * nta: added nta_sip_is_internal(). Tue Nov 27 13:43:41 EST 2007 Pekka.Pessi@nokia.com * test_htable2.c: removed conversion warning Conversion of size_t to isize_t when compiled wihout --disable-size-compat. Wed Nov 21 11:16:04 EST 2007 Daniele Rondina <geaaru at gmail dot com> * nta.c: nta_outgoing_*create() now uses NTATAG_TPORT() even if NTATAG_DEFAULT_PROXY() is set Thu Nov 22 09:36:21 EST 2007 Pekka.Pessi@nokia.com * nta.c: always use tport from NTATAG_TPORT() (even if it is bad) Add tests, too. Thu Nov 22 10:01:33 EST 2007 Pekka.Pessi@nokia.com * nta: added nta_msg_is_internal(). Mon Nov 26 17:08:16 EST 2007 Pekka.Pessi@nokia.com * test_proxy.c: removed memory leaks Mon Nov 26 17:08:35 EST 2007 Pekka.Pessi@nokia.com * test_basic_call.c: fixed --no-proxy tests Mon Nov 26 17:12:27 EST 2007 Pekka.Pessi@nokia.com * sofia-sip/su_wait.h: added su_msg_new(), su_msg_send_to(), su_msg_deinitializer() Reduce overhead from message passing. Mon Nov 26 19:15:41 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * configure.ac: defining HAVE_SOFIA_HTTP Mon Nov 26 19:23:05 EST 2007 Pekka.Pessi@nokia.com * nua: moved message passing into nua_stack.c. Recfactored reference counting. This seems to fix the memory leak within 1.12.7. Wed Nov 28 10:15:07 EST 2007 Pekka.Pessi@nokia.com * test_su.c: removed calls of deprecated (and unimplemented) functions su_clone_pause()/su_clone_resume() Thu Nov 22 09:59:13 EST 2007 Pekka.Pessi@nokia.com * nua_stack.c: more logging on bad authentication Mon Nov 26 19:49:34 EST 2007 Pekka.Pessi@nokia.com * tport.h, tport.c: updated tport_create() docs Thu Nov 29 12:17:40 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * tport: collecting statistics Thu Nov 29 12:21:10 EST 2007 Pekka.Pessi@nokia.com * rules/sofia.am: defining INTERNAL_INCLUDES so it can be used from any subdir utils/Makefile.am, libsofia-sip-ua/nua/Makefile.am: using INCLUDES from sofia.am Thu Nov 29 13:03:44 EST 2007 Pekka.Pessi@nokia.com * nua: moved test_nua contents in ltlibraries Thu Nov 29 13:05:25 EST 2007 Pekka.Pessi@nokia.com * Makefile.am, configure.ac: added subdir tests for Check-based module tests Fri Nov 30 09:03:14 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * nua_session.c: avoid NULL nua_session_state_t pointer in nua_update_client_report Thanks to Fabio Margarido for reporting this problem. git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@6539 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-12-06 18:44:14 +00:00
tport_recv_bytes(self, n, n);
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
if (N == 0) {
/* outbound-10 section 3.5.1 - send pong */
if (self->tp_ping >= 4)
tport_tcp_pong(self);
return 1;
}
}
veclen = tport_recv_iovec(self, &self->tp_msg, iovec, N, 0);
if (veclen == -1)
return -1;
msg = self->tp_msg;
msg_set_address(msg, self->tp_addr, (socklen_t)(self->tp_addrlen));
n = su_vrecv(self->tp_socket, iovec, veclen, 0, NULL, NULL);
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
if (n == SOCKET_ERROR)
return tport_recv_error_report(self);
assert(n <= N);
sync to current darcs tree: Tue Nov 20 11:46:34 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * nua_register.c: silenced warning with URL_INIT_AS() on Sun CC Tue Nov 20 11:47:06 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * test_register.c: fixed authentication caching issue Tue Nov 20 12:09:19 EST 2007 Pekka.Pessi@nokia.com * tport: tport_name_is_resolved() now uses host_is_ip_address() Tue Nov 20 12:53:54 EST 2007 Pekka.Pessi@nokia.com * tport.c: fixed EXPENSIVE unresolved bug. Wed Nov 21 07:10:40 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * test_tport.c: skipping tests on TLS if send fails. Wed Nov 21 11:46:42 EST 2007 Pekka.Pessi@nokia.com * test_nta.c: removed dead code Thu Nov 22 08:42:14 EST 2007 Pekka.Pessi@nokia.com * test_nta.c: refactored client side tests Thu Nov 22 09:39:45 EST 2007 Pekka.Pessi@nokia.com * nta: added nta_sip_is_internal(). Tue Nov 27 13:43:41 EST 2007 Pekka.Pessi@nokia.com * test_htable2.c: removed conversion warning Conversion of size_t to isize_t when compiled wihout --disable-size-compat. Wed Nov 21 11:16:04 EST 2007 Daniele Rondina <geaaru at gmail dot com> * nta.c: nta_outgoing_*create() now uses NTATAG_TPORT() even if NTATAG_DEFAULT_PROXY() is set Thu Nov 22 09:36:21 EST 2007 Pekka.Pessi@nokia.com * nta.c: always use tport from NTATAG_TPORT() (even if it is bad) Add tests, too. Thu Nov 22 10:01:33 EST 2007 Pekka.Pessi@nokia.com * nta: added nta_msg_is_internal(). Mon Nov 26 17:08:16 EST 2007 Pekka.Pessi@nokia.com * test_proxy.c: removed memory leaks Mon Nov 26 17:08:35 EST 2007 Pekka.Pessi@nokia.com * test_basic_call.c: fixed --no-proxy tests Mon Nov 26 17:12:27 EST 2007 Pekka.Pessi@nokia.com * sofia-sip/su_wait.h: added su_msg_new(), su_msg_send_to(), su_msg_deinitializer() Reduce overhead from message passing. Mon Nov 26 19:15:41 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * configure.ac: defining HAVE_SOFIA_HTTP Mon Nov 26 19:23:05 EST 2007 Pekka.Pessi@nokia.com * nua: moved message passing into nua_stack.c. Recfactored reference counting. This seems to fix the memory leak within 1.12.7. Wed Nov 28 10:15:07 EST 2007 Pekka.Pessi@nokia.com * test_su.c: removed calls of deprecated (and unimplemented) functions su_clone_pause()/su_clone_resume() Thu Nov 22 09:59:13 EST 2007 Pekka.Pessi@nokia.com * nua_stack.c: more logging on bad authentication Mon Nov 26 19:49:34 EST 2007 Pekka.Pessi@nokia.com * tport.h, tport.c: updated tport_create() docs Thu Nov 29 12:17:40 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * tport: collecting statistics Thu Nov 29 12:21:10 EST 2007 Pekka.Pessi@nokia.com * rules/sofia.am: defining INTERNAL_INCLUDES so it can be used from any subdir utils/Makefile.am, libsofia-sip-ua/nua/Makefile.am: using INCLUDES from sofia.am Thu Nov 29 13:03:44 EST 2007 Pekka.Pessi@nokia.com * nua: moved test_nua contents in ltlibraries Thu Nov 29 13:05:25 EST 2007 Pekka.Pessi@nokia.com * Makefile.am, configure.ac: added subdir tests for Check-based module tests Fri Nov 30 09:03:14 EST 2007 Pekka Pessi <Pekka.Pessi@nokia.com> * nua_session.c: avoid NULL nua_session_state_t pointer in nua_update_client_report Thanks to Fabio Margarido for reporting this problem. git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@6539 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-12-06 18:44:14 +00:00
tport_recv_bytes(self, n, n);
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
/* Check if message contains only whitespace */
/* This can happen if multiple PINGs are received at once */
if (initial) {
size_t i = ws_span(iovec->siv_base, iovec->siv_len);
if (i + self->tp_ping >= 4)
tport_tcp_pong(self);
else
self->tp_ping += (unsigned short)i;
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
if (i == iovec->siv_len && veclen == 1) {
SU_DEBUG_7(("%s(%p): received %u bytes of keepalive\n",
__func__, (void *)self, (unsigned)i));
msg_destroy(self->tp_msg), self->tp_msg = NULL;
return 1;
}
}
/* Write the received data to the message dump file */
if (self->tp_master->mr_dump_file)
tport_dump_iovec(self, msg, n, iovec, veclen, "recv", "from");
2011-07-31 23:36:05 +00:00
if (self->tp_master->mr_capt_sock)
tport_capt_msg(self, msg, n, iovec, veclen, "recv");
/* Mark buffer as used */
Update sofia-sip from darcs: Mon May 14 12:43:07 EDT 2007 martti.mela@nokia.com * su_base_port.c: fixed a double free in su_base_port_start_shared(). Fri May 25 13:56:23 EDT 2007 Pekka.Pessi@nokia.com * soa: added SOATAG_ORDERED_USER(), SOATAG_REUSE_REJECTED(). Allow replacing existing m=lines. Sun May 27 14:52:13 EDT 2007 Pekka.Pessi@nokia.com * msg_parser.c: fixed bug #1726034 Mon May 28 04:57:08 EDT 2007 Pekka.Pessi@nokia.com * test_nth.c: using non-blocking connect in test program, too. Mon May 28 04:58:05 EDT 2007 Pekka.Pessi@nokia.com * su.c: making all sockets non-blocking by default. Mon May 28 04:59:28 EDT 2007 Pekka.Pessi@nokia.com * m4/sac-su.m4: moved contents into sac-s2.m4 Mon May 28 05:32:26 EDT 2007 Pekka.Pessi@nokia.com * RELEASE: updated. Wed May 30 10:37:53 EDT 2007 Pekka.Pessi@nokia.com * m4/sac-su2.m4: added configure option --disable-tag-cast. Added SU_INLINE_TAG_CAST into sofia-sip/su_configure.h{,.in}. Using SU_INLINE_TAG_CAST in sofia-sip/sip_tag.h{,.in} sofia-sip/http_tag.h{,.in} sofia-sip/su_tag.h sofia-sip/su_tag_io.h sofia-sip/auth_module.h sofia-sip/nth_tag.h sofia-sip/nua_tag.h Fri Jun 1 15:11:52 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed tport_set_params() with secondary transports Fri Jun 1 15:13:23 EDT 2007 Pekka.Pessi@nokia.com * tport_type_tcp.c: checking for end-of-stream even if su_getmsgsize() promised more data Fri Jun 1 15:15:34 EDT 2007 Pekka.Pessi@nokia.com * tport: added tport_is_clear_to_send(), allow use of tport_pending() without msg The error callback from tport can now be registered even if there is no request pending on transport (e.g., when keeping a transport connection open for inbound messages). Fri Jun 1 15:16:43 EDT 2007 Pekka.Pessi@nokia.com * nta: not retrying after an transport error if application provided the transport Fri Jun 1 15:17:23 EDT 2007 Pekka.Pessi@nokia.com * sip: do not accept empty URIs (<>) in From, To, and other headers expecting name-addr format Fri Jun 1 15:17:43 EDT 2007 Pekka.Pessi@nokia.com * torture_url.c: added test for parsing empty URLs. Fri Jun 1 15:19:27 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.[hc]: use registered connections for outbound with TCP. Added test_proxy_close_tports() used testing recovering from TCP failures. Fri Jun 1 15:20:33 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: added --print-tags and --tags-a, --tags=b and --tags=c options Added more functions for handling events Fri Jun 1 15:22:08 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed some tests depending on delivery of responses in correct order Reordering might happen if some messages are sent over TCP, other over UDP. Fri Jun 1 15:27:55 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: re-registering in case the TCP connection towards proxy is closed In test_nua, Mr. B is now using TCP with the test proxy. Fri Jun 1 15:35:39 EDT 2007 Pekka.Pessi@nokia.com * nua/test_refer.c: fixed SIP payload checks Fri Jun 1 15:36:08 EDT 2007 Pekka.Pessi@nokia.com * nta_internal.h: added orq_user_tport field Tue Jun 5 06:16:43 EDT 2007 Pekka.Pessi@nokia.com * hide_email.sh: now fixing links, too. Fri Jun 15 05:34:29 EDT 2007 Pekka.Pessi@nokia.com * nua: fixed documentation entries for API functions left out from doxygen Thanks for Jerry Ricahrds for pointing this out. Wed May 23 10:26:26 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * Correct documentation for parameter type of NUTAG_WITH_SAVED Mon Jun 18 12:34:58 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * Make nua_saved_event_request() and hence NUTAG_WITH_SAVED resilient to a NULL event content git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5413 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-06-20 10:41:15 +00:00
msg_recv_commit(msg, n, n == 0);
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
if (n > 0)
self->tp_ping = 0;
Update sofia-sip from darcs: Mon May 14 12:43:07 EDT 2007 martti.mela@nokia.com * su_base_port.c: fixed a double free in su_base_port_start_shared(). Fri May 25 13:56:23 EDT 2007 Pekka.Pessi@nokia.com * soa: added SOATAG_ORDERED_USER(), SOATAG_REUSE_REJECTED(). Allow replacing existing m=lines. Sun May 27 14:52:13 EDT 2007 Pekka.Pessi@nokia.com * msg_parser.c: fixed bug #1726034 Mon May 28 04:57:08 EDT 2007 Pekka.Pessi@nokia.com * test_nth.c: using non-blocking connect in test program, too. Mon May 28 04:58:05 EDT 2007 Pekka.Pessi@nokia.com * su.c: making all sockets non-blocking by default. Mon May 28 04:59:28 EDT 2007 Pekka.Pessi@nokia.com * m4/sac-su.m4: moved contents into sac-s2.m4 Mon May 28 05:32:26 EDT 2007 Pekka.Pessi@nokia.com * RELEASE: updated. Wed May 30 10:37:53 EDT 2007 Pekka.Pessi@nokia.com * m4/sac-su2.m4: added configure option --disable-tag-cast. Added SU_INLINE_TAG_CAST into sofia-sip/su_configure.h{,.in}. Using SU_INLINE_TAG_CAST in sofia-sip/sip_tag.h{,.in} sofia-sip/http_tag.h{,.in} sofia-sip/su_tag.h sofia-sip/su_tag_io.h sofia-sip/auth_module.h sofia-sip/nth_tag.h sofia-sip/nua_tag.h Fri Jun 1 15:11:52 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed tport_set_params() with secondary transports Fri Jun 1 15:13:23 EDT 2007 Pekka.Pessi@nokia.com * tport_type_tcp.c: checking for end-of-stream even if su_getmsgsize() promised more data Fri Jun 1 15:15:34 EDT 2007 Pekka.Pessi@nokia.com * tport: added tport_is_clear_to_send(), allow use of tport_pending() without msg The error callback from tport can now be registered even if there is no request pending on transport (e.g., when keeping a transport connection open for inbound messages). Fri Jun 1 15:16:43 EDT 2007 Pekka.Pessi@nokia.com * nta: not retrying after an transport error if application provided the transport Fri Jun 1 15:17:23 EDT 2007 Pekka.Pessi@nokia.com * sip: do not accept empty URIs (<>) in From, To, and other headers expecting name-addr format Fri Jun 1 15:17:43 EDT 2007 Pekka.Pessi@nokia.com * torture_url.c: added test for parsing empty URLs. Fri Jun 1 15:19:27 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.[hc]: use registered connections for outbound with TCP. Added test_proxy_close_tports() used testing recovering from TCP failures. Fri Jun 1 15:20:33 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: added --print-tags and --tags-a, --tags=b and --tags=c options Added more functions for handling events Fri Jun 1 15:22:08 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed some tests depending on delivery of responses in correct order Reordering might happen if some messages are sent over TCP, other over UDP. Fri Jun 1 15:27:55 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: re-registering in case the TCP connection towards proxy is closed In test_nua, Mr. B is now using TCP with the test proxy. Fri Jun 1 15:35:39 EDT 2007 Pekka.Pessi@nokia.com * nua/test_refer.c: fixed SIP payload checks Fri Jun 1 15:36:08 EDT 2007 Pekka.Pessi@nokia.com * nta_internal.h: added orq_user_tport field Tue Jun 5 06:16:43 EDT 2007 Pekka.Pessi@nokia.com * hide_email.sh: now fixing links, too. Fri Jun 15 05:34:29 EDT 2007 Pekka.Pessi@nokia.com * nua: fixed documentation entries for API functions left out from doxygen Thanks for Jerry Ricahrds for pointing this out. Wed May 23 10:26:26 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * Correct documentation for parameter type of NUTAG_WITH_SAVED Mon Jun 18 12:34:58 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * Make nua_saved_event_request() and hence NUTAG_WITH_SAVED resilient to a NULL event content git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5413 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-06-20 10:41:15 +00:00
return n != 0;
}
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
/** Send to stream */
ssize_t tport_send_stream(tport_t const *self, msg_t *msg,
msg_iovec_t iov[],
size_t iovused)
{
#if __sun__ /* XXX - there must be a better way... */
if (iovused > 16)
iovused = 16;
#endif
return su_vsend(self->tp_socket, iov, iovused, MSG_NOSIGNAL, NULL, 0);
}
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
/** Calculate timeout if receive is incomplete. */
int tport_next_recv_timeout(tport_t *self,
su_time_t *return_target,
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
char const **return_why)
{
unsigned timeout = self->tp_params->tpp_timeout;
if (timeout < INT_MAX) {
/* Recv timeout */
if (self->tp_msg) {
su_time_t ntime = su_time_add(self->tp_rtime, timeout);
if (su_time_cmp(ntime, *return_target) < 0)
*return_target = ntime, *return_why = "recv timeout";
}
#if 0
/* Send timeout */
if (tport_has_queued(self)) {
su_time_t ntime = su_time_add(self->tp_stime, timeout);
if (su_time_cmp(ntime, *return_target) < 0)
*return_target = ntime, *return_why = "send timeout";
}
#endif
}
return 0;
}
/** Timeout timer if receive is incomplete */
void tport_recv_timeout_timer(tport_t *self, su_time_t now)
{
unsigned timeout = self->tp_params->tpp_timeout;
if (timeout < INT_MAX) {
if (self->tp_msg &&
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
su_time_cmp(su_time_add(self->tp_rtime, timeout), now) < 0) {
msg_t *msg = self->tp_msg;
msg_set_streaming(msg, (enum msg_streaming_status)0);
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
msg_set_flags(msg, MSG_FLG_ERROR | MSG_FLG_TRUNC | MSG_FLG_TIMEOUT);
tport_deliver(self, msg, NULL, NULL, now);
self->tp_msg = NULL;
}
#if 0
/* Send timeout */
if (tport_has_queued(self) &&
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
su_time_cmp(su_time_add(self->tp_stime, timeout), now) < 0) {
stime = su_time_add(self->tp_stime, self->tp_params->tpp_timeout);
if (su_time_cmp(stime, target) < 0)
target = stime;
}
#endif
}
}
/** Calculate next timeout for keepalive */
int tport_next_keepalive(tport_t *self,
su_time_t *return_target,
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
char const **return_why)
{
/* Keepalive timer */
unsigned timeout = self->tp_params->tpp_keepalive;
if (timeout != 0 && timeout != UINT_MAX) {
if (!tport_has_queued(self)) {
su_time_t ntime = su_time_add(self->tp_ktime, timeout);
if (su_time_cmp(ntime, *return_target) < 0)
*return_target = ntime, *return_why = "keepalive";
}
}
timeout = self->tp_params->tpp_pingpong;
if (timeout != 0) {
if (self->tp_ptime.tv_sec && !self->tp_recv_close) {
su_time_t ntime = su_time_add(self->tp_ptime, timeout);
if (su_time_cmp(ntime, *return_target) < 0)
*return_target = ntime, *return_why = "waiting for pong";
}
}
return 0;
}
/** Keepalive timer. */
void tport_keepalive_timer(tport_t *self, su_time_t now)
{
unsigned timeout = self->tp_params->tpp_pingpong;
if (timeout != 0) {
if (self->tp_ptime.tv_sec && !self->tp_recv_close &&
su_time_cmp(su_time_add(self->tp_ptime, timeout), now) < 0) {
SU_DEBUG_3(("%s(%p): %s to " TPN_FORMAT "%s\n",
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
__func__, (void *)self,
"closing connection", TPN_ARGS(self->tp_name),
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
" because of PONG timeout"));
tport_error_report(self, EPIPE, NULL);
if (!self->tp_closed)
tport_close(self);
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
return;
}
}
timeout = self->tp_params->tpp_keepalive;
if (timeout != 0 && timeout != UINT_MAX) {
if (su_time_cmp(su_time_add(self->tp_ktime, timeout), now) < 0) {
tport_tcp_ping(self, now);
}
}
}
/** Send PING */
int tport_tcp_ping(tport_t *self, su_time_t now)
{
ssize_t n;
char *why = "";
if (tport_has_queued(self))
return 0;
n = send(self->tp_socket, "\r\n\r\n", 4, 0);
if (n > 0)
self->tp_ktime = now;
if (n == 4) {
if (self->tp_ptime.tv_sec == 0)
self->tp_ptime = now;
}
else if (n == -1) {
int error = su_errno();
why = " failed";
if (!su_is_blocking(error))
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
tport_error_report(self, error, NULL);
else
why = " blocking";
return -1;
}
SU_DEBUG_7(("%s(%p): %s to " TPN_FORMAT "%s\n",
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
__func__, (void *)self,
"sending PING", TPN_ARGS(self->tp_name), why));
return n == -1 ? -1 : 0;
}
/** Send pong */
int tport_tcp_pong(tport_t *self)
{
self->tp_ping = 0;
if (tport_has_queued(self) || !self->tp_params->tpp_pong2ping)
return 0;
SU_DEBUG_7(("%s(%p): %s to " TPN_FORMAT "%s\n",
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
__func__, (void *)self,
"sending PONG", TPN_ARGS(self->tp_name), ""));
return send(self->tp_socket, "\r\n", 2, 0);
}
/** Calculate next timer for TCP. */
int tport_tcp_next_timer(tport_t *self,
su_time_t *return_target,
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
char const **return_why)
{
return
Merge up to the most recent sofia-sip darcs tree. Includes the following patches from darcs: Tue Aug 21 09:38:59 EDT 2007 Pekka.Pessi@nokia.com * tport_type_udp.c: checking error while checking that MSG_TRUNC works. Shall I pull this patch? (1/43) [ynWvpxqadjk], or ? for help: y Tue Aug 21 10:49:33 EDT 2007 Pekka.Pessi@nokia.com * nua_params.c: NUTAG_SIPS_URL() now sets the handle target, too. Problem reported by Jari Tenhunen. Shall I pull this patch? (2/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 11:22:42 EDT 2007 Pekka.Pessi@nokia.com * nta.c: do not destroy INVITE transaction if it has been CANCELed Handle gracefully cases where the INVITE transaction is destroyed immediately after canceling it. The old behaviour was to left it up to the application to ACK the final response returned to INVITE. Thanks for Fabio Margarido for reporting this problem. Shall I pull this patch? (3/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 13:02:01 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added test with user SDP containing already rejected media Shall I pull this patch? (4/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com * nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (5/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:41:20 EDT 2007 Pekka.Pessi@nokia.com UNDO: nta: added option for processing orphan responses matching with a dialog The orphan responses matching with the dialog can now be processed by the response callback.The dialog leg can be created with NTATAG_RESPONSE_CALLBACK() or a response callback can be later bound to the leg with nta_leg_bind_response(). This is practically useful only with 200 OK responses to the INVITE that are retransmitted by the UAS. By default, the retransmission are catched by the ACK transaction (which then retransmits the ACK request message). However, after ACK transaction times out, the retransmitted 200 OK indicates most probably that the ACK request messages do not reach UAS. Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (6/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:00:10 EDT 2007 Pekka.Pessi@nokia.com * nta.c: disabled nta_msg_ackbye(). Fix for sf.net bug #1750691 Thanks for Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (7/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 06:54:38 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for sf.net bug #1750691 Shall I pull this patch? (8/43) [ynWvpxqadjk], or ? for help: y Thu Aug 30 07:03:45 EDT 2007 Pekka.Pessi@nokia.com * test_nua: added test for nua_bye() sending CANCEL Shall I pull this patch? (9/43) [ynWvpxqadjk], or ? for help: y Fri Aug 31 12:08:09 EDT 2007 Pekka.Pessi@nokia.com * url.c: fixed escaping of '/' %2F, ';' %3B and '=' %3D in URL path/params Thanks for Fabio Margarido for reporting this bug. Shall I pull this patch? (10/43) [ynWvpxqadjk], or ? for help: y Mon Sep 3 10:14:55 EDT 2007 Pekka.Pessi@nokia.com * url.c: do not un-escape %40 in URI parameters. Do not unescape %2C, %3B, %3D, or %40 in URI parameters, nor %2C, %2F, %3B, %3D, or %40 in URI path. The @ sign can be ambiguous in the SIP URL, e.g., <sip:test.info;p=value@test.com> can be parsed in two ways: 1) username contains test.info;param=value and host part has test.com 2) empty username, host part test.info, URI parameter p=value@test.com Previously Sofia URL parser converted escaped '@' at signs (%40) in the URI parameters to the unescaped form. The resulting URI could be ambiguous and sometimes fail the syntax check if there was no '@' sign before the unescaped one. Thanks for Jan van den Bosch and Mikhail Zabaluev for reporting this bug. Shall I pull this patch? (11/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 04:59:57 EDT 2007 Pekka.Pessi@nokia.com * tport.c: fixed indenting, logging Shall I pull this patch? (12/43) [ynWvpxqadjk], or ? for help: y Fri Jul 13 12:47:33 EDT 2007 Pekka.Pessi@nokia.com * nua/test_proxy.h, nua/test_proxy.c: added support for multiple domains Each domain has its own registrar and authentication module. Shall I pull this patch? (13/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:19:33 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: added timestamp to event logging Shall I pull this patch? (14/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:20:12 EDT 2007 Pekka.Pessi@nokia.com * test_nua: fixed timing problems in testing. Shall I pull this patch? (15/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:04 EDT 2007 Pekka.Pessi@nokia.com * test_ops.c: reduce su_root_step() delay to 0.1 seconds Shall I pull this patch? (16/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:31:22 EDT 2007 Pekka.Pessi@nokia.com * test_register.c: fixed timing problem Shall I pull this patch? (17/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 17:03:46 EDT 2007 Pekka.Pessi@nokia.com * test_100rel.c: fixed timing problems resulting in events being reordered Shall I pull this patch? (18/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:40:53 EDT 2007 Pekka.Pessi@nokia.com * nua (test_init.c, test_register.c): using test_proxy domains Shall I pull this patch? (19/43) [ynWvpxqadjk], or ? for help: y Thu Aug 23 12:12:32 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: added cleanup code Shall I pull this patch? (20/43) [ynWvpxqadjk], or ? for help: y Fri Aug 24 09:35:35 EDT 2007 Pekka.Pessi@nokia.com * nta.c: increase lifetime of ACK transaction from T4 to T1 x 64 nta.c creates a ACK transaction in order to restransmit ACK requests when ever a retransmitted 2XX response to INVITE is received. The UAS retransmits the 2XX responses for 64 x T1 (32 second by default). Partially fixes the sf.net bug #1750691 reported by Mikhail Zabaluev. Shall I pull this patch? (21/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 10:21:04 EDT 2007 Pekka.Pessi@nokia.com * Makefile.am: generating libsofia-sip-ua/docs/Doxyfile.rfc before making manpages Shall I pull this patch? (22/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:05:33 EDT 2007 Pekka.Pessi@nokia.com * sofia-sip/tport_tag.h: added TPTAG_KEEPALIVE(), TPTAG_PINGPONG(), TPTAG_PONG2PING() Shall I pull this patch? (23/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:09:06 EDT 2007 Pekka.Pessi@nokia.com * tport: added ping-pong keepalive on TCP. replaced single tick with connection-specific timer Now detecting closed connections on TLS, too. Added tests for idle timeout, receive timeout, ping-pong timeout. Shall I pull this patch? (24/43) [ynWvpxqadjk], or ? for help: y Fri Jul 6 10:19:32 EDT 2007 Pekka.Pessi@nokia.com * nta.c: added nta_incoming_received() Shall I pull this patch? (25/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 11:29:56 EDT 2007 Pekka.Pessi@nokia.com * nua_session.c: delay transition to ready when O/A is incomplete Delay sending ACK and subsequent transition of call to the ready state when the 200 OK response to the INVITE is received if the SDP Offer/Answer exchange using UPDATE/PRACK was still incomplete. Previously, if the O/A using UPDATE or PRACK was incomplete and an 200 OK was received, the call setup logic regarded this as a fatal error and terminated the call. Thanks for Mike Jerris for detecting and reporting this bug. Shall I pull this patch? (26/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:22:46 EDT 2007 Pekka.Pessi@nokia.com * test_call_reject.c: testing Retry-After Shall I pull this patch? (27/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:42:51 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using rudimentary outbound support in B's proxy. Shall I pull this patch? (28/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:48:33 EDT 2007 Pekka.Pessi@nokia.com * nua_register.c: added some logging to nua_register_connection_closed() Shall I pull this patch? (29/43) [ynWvpxqadjk], or ? for help: y Wed Jul 25 12:43:57 EDT 2007 Pekka.Pessi@nokia.com * test_nua: using AUTHTAG_MAX_NCOUNT(1) for Mr. C C is now challenged every time. Shall I pull this patch? (30/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 11:05:19 EDT 2007 Pekka.Pessi@nokia.com * nua/test_100rel.c: fixed timing problem re response to PRACK and ACK Shall I pull this patch? (31/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 06:02:50 EDT 2007 Mikhail Zabaluev <mikhail.zabaluev@nokia.com> * DIST_SUBDIRS must include everything unconditionally Shall I pull this patch? (32/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:53:04 EDT 2007 Pekka.Pessi@nokia.com * test_soa.c: silenced warnings Shall I pull this patch? (33/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com * nua: refactored dialog refresh code Shall I pull this patch? (34/43) [ynWvpxqadjk], or ? for help: y Mon Jul 23 16:59:48 EDT 2007 Pekka.Pessi@nokia.com UNDO: nua: refactored dialog refresh code Shall I pull this patch? (35/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:01:25 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc]: renamed functions setting refresh interval Shall I pull this patch? (36/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:15:03 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.[hc], nua_stack.c: added nua_dialog_repeat_shutdown() Shall I pull this patch? (37/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:19:20 EDT 2007 Pekka.Pessi@nokia.com * nua_dialog.h: renamed nua_remote_t as nua_dialog_peer_info_t Shall I pull this patch? (38/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:23:04 EDT 2007 Pekka.Pessi@nokia.com * nua_stack.c: added timer to client request in order to implement Retry-After Shall I pull this patch? (39/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 12:33:53 EDT 2007 Pekka.Pessi@nokia.com * nua: added backpointers to nua_dialog_usage_t and nua_dialog_state_t Shall I pull this patch? (40/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 13:56:48 EDT 2007 Pekka.Pessi@nokia.com * test_nua.c: abort() in timeout alarm function if -a is given Shall I pull this patch? (41/43) [ynWvpxqadjk], or ? for help: y Thu Sep 6 17:13:18 EDT 2007 Pekka.Pessi@nokia.com * nua_subnotref.c: include SIPTAG_EVENT() in the nua_i_notify tag list Shall I pull this patch? (42/43) [ynWvpxqadjk], or ? for help: y Mon Sep 10 12:27:53 EDT 2007 Pekka.Pessi@nokia.com * nua: save Contact from target refresh request or response. Save the Contact header which the application has added to the target refresh requests or responses and use the saved contact in subsequent target refresh requests or responses. Previously the application had no way of specifying the Contact included in the automatic responses to target refresh requests. Thanks for Anthony Minessale for reporting this problem. Shall I pull this patch? (43/43) [ynWvpxqadjk], or ? for help: y git-svn-id: http://svn.freeswitch.org/svn/freeswitch/trunk@5692 d0543943-73ff-0310-b7d9-9358b9ac24b2
2007-09-10 20:45:25 +00:00
tport_next_recv_timeout(self, return_target, return_why) |
tport_next_keepalive(self, return_target, return_why);
}
/** TCP timer. */
void tport_tcp_timer(tport_t *self, su_time_t now)
{
tport_recv_timeout_timer(self, now);
tport_keepalive_timer(self, now);
tport_base_timer(self, now);
}