[omniORB] surviving thread on 3.04

[Renzo Tomaselli]

Hi all,
    while hunting for a strange core occurring on Linux *after* calling
orb->shutdown(), I'm noticing a thread which survives after the above call.
This thread has been observed on a NT platform, using OmniORB 3.04 out of
the box. Because of the reasons below, it certainly occurs on any platform
as well.
This thread manages the POA etherealization_queue and its code is in
taskqueue.cc. I can see from there as a die() method was provided to kill
this thread, but actually it is never called from anywhere.
I guess is missing from omniOrbPOA::shutdown().
While I still don't know whether this relates to our exit core (we'll
investigate further), I believe this thread should go away before exiting as
a matter of a fair shutdown.
Thanks,
                                             Renzo Tomaselli
---------------------------------------------------------------------------
TecnoTP s.n.c. Special Information System Design
Maso Pelauchi I38050 Ronchi Valsugana,  Trento TN  ITALY
Tel. +39 0461 773164      Fax. +39 0461 771514
e-mail: renzo.tomaselli@tecnotp.it
---------------------------------------------------------------------------