Re: [Chicken-users] long pointer equivalent

[Peter Keller]

On Fri, Aug 23, 2002 at 09:02:34AM +0200, felix wrote:
> >
> > So, what is the verdict about this?
> 
> Opps, Sorry. I'm a bit slow these days (lots of work).
> So what do you need:

> 1) s64vector/u64vector-support for SRFI-4
Don't need this..

> 2) A foreign type ìnt64/unsigned-int64
Don't quite need this.

3) A foreign type for long pointer/unsigned long pointer.
This is what I need.

See, you can't automatically assume long is 64 bits. Sometimes it is,
sometimes it isn't. 

Although, I could very well be misunderstanding the solution you've given
and what you've said just might be perfect. 

> There are some minor troubles that make this not quite trivial
> (it has to compile at least on gcc and msvc; and the handling of
> those data types needs some special work when transforming
> arguments/results to foreign code. A 64 bit integer would necessarily
> be passed as a float, since I don't want to add another basic
> numeric type - that would need a complete re-write of all
> procedures that handle numbers)

ok.

-pete