[omniORB] "multiplexed calls" in local case?

[Duncan Grisby]

On Monday 26 November, tihomir.sokcevic@acterna.com wrote:

> I wrote a little echo-like example to test the "multiplexed calls"
> ability of omniORB4.
>
> I got this results:
>   different machines -> everything works fine,
>   different processes on one machine -> everything works fine,
>   different threads in one process -> omniORB4 deadlocks!
> 
> Is this a bug or a feature (a question of speed)?

I didn't completely follow what your example is doing, but I assume it
relies on the fact that the client's invocation can return before the
operation is dispatched to the servant. This doesn't happen in the
colocated case, since the same thread that did the client-side
invocation does the servant up-call. The spec doesn't require anything
much about oneways, so this is perfectly compliant.

Basically, oneways should be used with _extreme_ caution.

If that isn't what you're seeing, please can you post a full example
program.

Cheers,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --