[omniORB] GIOP reply status error when using a temporarly blocking thread

Laurent Pointal pointal@lure.u-psud.fr
Thu, 26 Aug 1999 14:30:27 +0200


[carreful when replying, I sent the mail to both Fnorb and OmniORB2 lists]

Platform: WinNT 4+SP1
Server: MSVC++ 5+SP3, OmniOrb 2.7.1.
Client: Python 1.5.2, Fnorb1.01.

I have been able to interconnect my C++ server to my Python client. It
works well when the server thread directly process the request then return.
But when the server thread has to delay the processing to another thread
and wait for a signal from this other thread, all seems to be good in the
server side (I can trace execution in my log), but the return from the
client call fail in exception when it tests the GIOP reply header status
and find a SYSTEM_EXCEPTION (this is the Fnorb  GIOPClient.py line 430 test
which fail).


Dont know exactly what are the GIOP protocols, but there may be a mismatch
between omniORB2 and Fnorb... have you an idea?

Thanks.

Laurent.

[note: for my previous problem (multiple defined _DllMain symbol between
mfcs40d.lib and MSVCRTD.lib when linking), its a known problem of link
order of Microsoft libs and can be resolved, see Knowledge Base Q148652 &
Q72651.]


Laurent POINTAL - CNRS/LURE - Service Informatique Experiences
Tel/fax: 01 64 46 82 80 / 01 64 46 41 48
email  : pointal@lure.u-psud.fr  ou  lpointal@planete.net 
ouebe  : http://www.planete.net/~lpointal