Thanks Juha,
sems ser_rel_0_9_0 should work fine with openser 1.0.1 and svn trunk version of sems should work ok with cvs head version of openser.
As you suggested, I upgraded to the CVS head of OpenSER to match the SVN head of SEMS. But SEMS still bails out when it replies, see attached details. (Not sure where to ask, but it seems that SER is not picking up where SEMS left off.)
I cannot figure out what the connection between SEMS' reply socket and SER is. And that appears to cause the problems. Documentation is rather thin on this issue.
In openser.cfg I have set
unix_sock="/tmp/openser.sock"
...and I call...
t_write_unix ("/tmp/sems.sock", "echo")
In sems.conf I have
socket_name=/tmp/sems.sock reply_socket_name=/tmp/sems.reply.sock
The latter name is not connected to anything of SER, as far as I can tell. If I set reply_socket_name to /tmp/openser.sock I get the same crash as below. If I set reply_socket_name to /tmp/sems.sock there is no crash but also no behaviour.
In all cases, SER seems to be doing a lot of reply resending amd cleanup.
Thanks for any suggestions you may have,
Cheers,
Rick van Rein.
Below is the debugging output that ends in a crash. The core dump confirms the function names that are in the output.
------------ 8< ------------ 8< ------------ 8< ------------ 8< ------------
(79899) DEBUG: write_to_socket (AmUtils.cpp:587): sending: <:t_reply:/tmp/167EB0 E7-443E3484000AE409-0884E000 200 OK 31998:80608280 41C67EA6-443E3484000A97C9-08103000 Supported: timer Session-Expires: 10;refresher=uas Contact: sip:vanrein@10.0.0.10 Content-Type: application/sdp . v=0 o=username 0 0 IN IP4 127.0.0.1 s=session c=IN IP4 127.0.0.1 t=0 0 m=audio 10000 RTP/AVP 0 a=rtmap:101 telephone-event/8000 a=fmtp:101 0-15 a=direction:passive .
(79899) DEBUG: write_to_socket (AmUtils.cpp:613): write to unix socket: complete d Bus error (core dumped)