Commits
Mike Jerris committed 274a956f00f
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