Re: [orbitcpp-list] Re: cpp branch: CORBA_Object struct hidden?

[Michael Meeks <michael ximian com>]

Hi Murray,

On Tue, 2002-02-19 at 09:58, Murray Cumming wrote:
> So CORBA::Object would inherit from ORBITCppInternalStruct?

	Nope; the C++ wrapper would just keep a pointer to that.

> Then I don't
> see the purpose of cpp_impl_pointer, and I don't see how this is
> different than having the CORBA_Object as a member field instead of
> inheriting.

	It isn't much different.

> There's no publicness problem in having the CORBA_Object as a member. 

	Quite.

> The current problem is that there is no way to get an already existing
> C++ wrapper instance from a C CORBA_Object instance. It's the lack of a
> g_object_set_qdata() as mentioned in another thread.

	Does that mean that you want to be able to call C++ wrapper methods on
object clients instantiated from C ? I suppose that's prolly necessary
due to the way is_equivalent works etc. hmm; in which case we need to
add a user_data pointer to the CORBA_Object - go for it :-) We could use
a qdata if you want.

> Maybe there's a better way to do this. After all, how would a Perl or
> Python binding libbonobo binding create a Perl/Python stub for an
> existing C CORBA_Object instance?

	Fair enough.

> For instance, is there any kind of
> instance ID? Can I create a 2nd CORBA_Object stub instance for the same
> BonoboWindow?

	Whatever; I really don't quite see why you need to associate data with
the client object handle; it seems somewhat strange - no other client
code does this from C - what are you going to use it for ? Associating
stuff with the server I can see would be useful but ...

	Regards,

		Michael.

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot