corba and scripting questions

[Martijn van Beers <sauron charm il fontys nl>]

Hi,

I've been redesigning the IDL for photon, and I've run into a few problems
I have the following structure

there's a Part interface, which is the base for each part of photon.
then each part has it's own interface, which derives from part

the problem is that since I'm not going to implement all the parts myself,
I put the Part interface in it's own file. Each part has a .idl file as
well, which includes the Part .idl.

When I run these files through orbit-idl, it generates code for Part in every
output file set. So, when I would try and compile the project-management
part, which must talk to every part, I'd have a lot of functions doubled.

Is there any way to solve this (except by putting it all in one big idl file)?

Also, I have an interface that represents a 'target', which should be shared
among several parts, so it's objref gets passed as a param in some functions.
This gives approximately the same problem as above, but can be avoided by
passing it as a generic CORBA_Object, right?


I intend to have the project-management part of photon get/put all it's
information from configure.in/Makefile.am, so I don't get a lot of duplication
(and thus possible synchronisation problems). The question is how to do the
parsing. Could I re-use automake/autoconf itself? If not, I was thinking of
writing some perl thing. But since scheme has been declared the gnome scripting
language of choice, I was wondering if it would be more difficult to do it in
scheme.

Martijn


-- 
Martijn van Beers
martijn@earthling.net