[jcifs] Re: jarapac, idlc, and TNG

[Elrond]

On Mon, Sep 27, 2004 at 07:24:06PM -0400, Michael B Allen wrote:
> Elrond said:
> > Also I'm quite undecided, if the emited code should do
> > marshallng and unmarshalling in one function (which current
> 
> Well I'll write the skeleton and leave it to you to play with. If your
> code isn't really generic that's ok. I only ask that you stick to strictly
> confomant MIDL compatible IDL [1]. We want to increase demand for it so we
> can use it too.

That's _exactly_, what I want!


> If you end up with [value(r->len)] PIDL kinds of stuff
> your on your own :-)

If I wanted pidl, I'd be using pidl...


> > [...]
> >> It's eric.glassATgm ail.com.
> >
> > Is he on jcifs? if so, he might already have read this.
> 
> Yes but he has a tendency to disappear for a while sometimes so ...

*g* That might be said about me too...
(Not that I'm really off, just not really reading the MLs)

I'll drop him a personal note.

> Mike
> 
> [1] Although the Java emitter currently supports an optional operation
> numer [op(0x01)] attribute that is not supported by MIDL. I could just not
> bear to have a bunch of dummy functions.

_That_'s a useful addition.

To be honest, I don't really want to mess with the idl
parser at all. I just want to write the emitter for C.

And I surely will take a look at the java emitter a lot.
(In fact, I already did and started rewriting current tng's
marshalling code in response to seeing it.)

And I hope, we will be able to share code for both
emitters.


    Elrond