Hi, On 07/01/2015 11:46 AM, Stefan Sayer wrote: > you're right. AmOfferAnswer::onReplyOut should not necessarily try to > get an SDP body in that case and may just send out the 200 without > SDP, or put in the SDP from the 183. > > The proper solution to this is to add another state > OA_PreviewCompleted which deals with the "preview" that may be > received here, as described in > https://tools.ietf.org/html/rfc6337#section-3.1.1 > In checkStateChange, the transition from PreviewCompleted to Completed > can then be ignored.
a similar issue affects me, using version 1.6.1 + a few patches In my case I have INVITE w/SDP, reliable 183 with SDP, PRACK, 180 w/o SDP, 200 OK w/SDP. After 200 OK sems gets an error and sends a BYE: > Dec 17 20:47:54 sipwise ngcp-sems[4983]: [#7fdeb2dad700] [relaySip, > AmB2BSession.cpp:862] DEBUG: relaying SIP reply 200 Ok > Dec 17 20:47:54 sipwise ngcp-sems[4983]: [#7fdeb2dad700] [reply, > AmBasicSipDialog.cpp:583] DEBUG: reply: transaction found! > Dec 17 20:47:54 sipwise ngcp-sems[4983]: [#7fdeb2dad700] [getSdpBody, > AmOfferAnswer.cpp:478] DEBUG: No SDP Offer. > Dec 17 20:47:54 sipwise ngcp-sems[4983]: [#7fdeb2dad700] [reply, > AmBasicSipDialog.cpp:600] DEBUG: onTxReply failed > Dec 17 20:47:54 sipwise ngcp-sems[4983]: [#7fdeb2dad700] [relaySip, > AmB2BSession.cpp:872] ERROR: dlg->reply() failed > Dec 17 20:47:54 sipwise ngcp-sems[4983]: [#7fdeb2dad700] [updateCallStatus, > CallLeg.cpp:1226] DEBUG: A leg 4EE66C79-5672AEF80000E9F5-E5402700 changing > status from Ringing to Disconnected > Dec 17 20:47:54 sipwise ngcp-sems[4983]: [#7fdeb2dad700] [terminateOtherLeg, > CallLeg.cpp:286] DEBUG: trying to terminate other leg in Disconnected state > -> terminating the others as well while technically one offer should have only one answer and 200 should be without SDP, I guess I'm going to hit this issue this or another way, right? Though I can't believe that 1.6.1 has broken PRACK support, so maybe something else in replies from the UA is wrong. Stefan, are you willing to have a look at the log off-list? Thanks, Andrew _______________________________________________ Semsdev mailing list Semsdev@lists.iptel.org http://lists.iptel.org/mailman/listinfo/semsdev
