[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Document the current module system?
From: |
NIIBE Yutaka |
Subject: |
Re: Document the current module system? |
Date: |
Mon, 23 Apr 2001 18:32:05 +0900 (JST) |
Marius Vollmer wrote:
> - A module might be provided completely by compiled code.
>
> - Compiled code might provide only part of a module.
>
> - Compiled code might want to provide bindings only on demand (to save
> initialization time) even if it has been linked already.
It reminds me an issue. Module written in C provides same interface
to Scheme world. A procedure or a variable can be implemented in C or
Scheme, and it can be accessed from Scheme world with no difference.
On the other hand, from C world, it's not that simple. If the
procedure has C binding, we can call it directly, but if not, we need
to use "apply" to invoke the procedure. I don't think it's good,
because programmer needs to know the implementation a priori.
It would be good if we could directly invoke the Scheme procedure
(with no apply). I think that it's possible for Guile to provide
auto-generated glue-code (stub), if needed.
If these scheme will be supported by Guile, a module always has
interface for C, even if it is written in Scheme.
It would be great if the language is just an implementation issue in
Guile.
--
Document the current module system?, NIIBE Yutaka, 2001/04/19
- Re: Document the current module system?, Martin Grabmueller, 2001/04/19
- Re: Document the current module system?, Rob Browning, 2001/04/20
- Re: Document the current module system?, NIIBE Yutaka, 2001/04/23
- Re: Document the current module system?, NIIBE Yutaka, 2001/04/23
- Re: Document the current module system?, Martin Grabmueller, 2001/04/23
- Recursive namespace, Keisuke Nishida, 2001/04/23
- Recursive namespace, NIIBE Yutaka, 2001/04/23
- Re: Recursive namespace, Keisuke Nishida, 2001/04/24