[omniORB] problem using naming service with bidir transport

[Duncan Grisby]

On Friday 17 September, Harri Pasanen wrote:

> This has the symptoms of a some sort of race condition.

Indeed. I've seen it occur once. As soon as I tried to figure out what
was going on, I never managed to make it happen again.

One thing that would help would be if you can manage to get a trace
with traceLevel 40 and also traceThreadId 1. It's very odd because the
client sends a GIOP 1.0 request, gets a GIOP 1.0 reply, and then blows
up in the GIOP 1.2 parser. It shouldn't even be thinking about GIOP
1.2. I can only think that the thread watching for bidirectional
requests picked up the reply and got confused about it, but that isn't
meant to happen. Seeing the thread ids will give more idea about that.

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --