[omniORB] servant destruction and ref counting

[Stefan Seefeld]

Duncan Grisby wrote:

> On Monday 30 April, Stefan Seefeld wrote:
> 
>> hmm, an easy applet I wrote doesn't show the behavior I'm trying to
>> eliminate.  What I'm seeing in my application is this:
>> 
>> omniORB: Deactivating: root/IDL:Warsaw/WidgetKit:1.0#4<16777216>
>> (has local refs).
> 
> 
> That's not an error of any sort. It's just an informational message
> that there were still active object references to the object when the
> servant was deactivated. It doesn't prevent the servant object from
> being deleted.
> 
> Local refs are any CORBA object references to the object in the same
> address space. If you think you have already released all references
> to the object, you must have a reference count leak. I'll think about
> adding the tracing you want, but it's a bit awkward since the
> reference counting is very much on the critical path.

figured. If it hurts performance, it'll be probably better to provide a 
binary that is compiled with
special preprocessor flags such as DEBUG.
Ok, let me sum up to see whether I understood well:

There are indeed local references indicating an error in reference 
counting (and that is
what I'm looking for). However, this shouldn't prevent the servant from 
being deleted.

Now, you didn't tell me what a reference in this context actually is.

Regards,    Stefan