[omniORB] servant destruction and ref counting

[Stefan Seefeld]

Duncan Grisby wrote:

> On Monday 23 April, Stefan Seefeld wrote:
> 
>> Now I'm a bit confused, as I always thought servant and reference
>> life cycles were completely independent, no matter whether they
>> are colocated or not.
> 
> 
> They should be.


that's good to hear. Otherwise I would be in a deadlock situation.

>  The presence of local object references shouldn't
> prevent a servant from being deactivated. Can you come up with a
> minimal program which exhibits the problem you are seeing?

ok, I'll try.

> What sort of access to reference counts do you think would be useful?


I'd like to be able to log all ref count activity, if it doesn't hurt 
performance (i.e. if not logging
would be as efficient as the current code).
I'd like to log all _duplicate/CORBA::release calls (with a 'this' 
pointer or some other identification
so I can grep for it), as well as _add_ref()/_remove_ref() (only in 
PortableServer::RefCountBase, of course).
I don't know at which traceLevel that fits in best, though.

Regards,    Stefan