D/BUS IDL compiler ...

Marco Canini marco.canini@fastwebnet.it
Wed, 10 Mar 2004 12:27:56 +0100

On Wed, 2004-03-10 at 12:03, Michael Meeks wrote:
> Hi Carlos,
> On Mon, 2004-03-08 at 09:33 +0200, Carlos.Guerreiro@nokia.com wrote:
> > Looks like Meeks is working on the IDL already ;)
> 	;-) although not so hard. Ultimately, I'm persuaded we have to have IDL
> as the authoritative, standard contract definitions, rather than parsing
> C headers (and/or perl,python,-whatever language the service is
> implemented in- to try and extract interface data / documentation from
> that).
I agree with you Michael,
the very first objective of a remote invocation framework is to define a
single format for exposing the interface.
That has to be language independent and mapped to specific language in a
second step.
IMHO RPC, CORBA and other successful IPC technologies are inspired by
the principle of design by contract theorized here
http://archive.eiffel.com/doc/manuals/technology/contract/ (to be honest
a part of it)
Thinking at it now that it's time to make decisions it could make sense
to add a form of preconditions to IDL (just a thought).

> 	Clearly my prototype compiler needs re-factoring though.
> > > 	* The license is LGPL/X11 - I'm still deeply concerned with the
> > > 	  unresolved AFL legal issue.
> > 
> > Do you know what is the exact legal issue he's worrying about?
> 	It's no issue if you use D/BUS under the GPL - although this may incur
> problems with embedded systems etc. (AFAIR). My problem (well not mine
> really, but...) is with section 10 of the AFL - which (it seems) might
> be construed to destroy the usefulness of a defensive patent portfolio -
> something vital to any large company. I believe I rambled in an
> ill-educated way in the past on the list about this - dig there. I
> understand Havoc is investigating / sorting the issue out though.
> 	Regards,
> 		Michael.
> -- 
>  michael@ximian.com  <><, Pseudo Engineer, itinerant idiot
Marco Canini <marco.canini@fastwebnet.it>