C interface (was: Elegant code)

Alex McDonald <blog@[EMAIL PROTECTED]
> writes:
>Then there is hope for "fat Forths" like Win32Forth. If there's one
>thing W32F encourages, it's getting leverage from existing libraries
>of code (DLLs primarily). The only downside at the moment is the
>automation of the interface, where MPE's VxForth has the neat ability
>to im****t C function headers directly. I've worked a little with SWIG
>(http://www.swig.org/)
and used XSLT on the XML output it generates,
>but it's far from automatic.
>
>Anton Ertl started some time back on a generalised C function
>interface (http://www.complang.tuwien.ac.at/papers/ertl06.ps.gz
and
>http://www.complang.tuwien.ac.at/papers/ertl07euroforth.ps.gz).
More
>please.

What has been described there is implemented in the current
development version (but it needs more poli****ng before we can release
it for use on machines that may have several Gforth processes).

This also processes C header files automatically.  However, in
addition to the C types, which are described in the C header file, we
need to know the Forth types of the function arguments.

One approach to this is to have a mapping from the basic C type to the
Forth type, but that's not ****table across platforms: on one platform
the C type of an argument can be a long long, which might be mapped to
a double-cell in Forth, and on another it can be a long, which may be
mapped to a single cell in Forth.

The current approach in development Gforth is to declare the Forth
types manually, but a student of mine works on a SWIG-based tool that
generates declarations automatically from C header files, based on a
mapping from the abstract C type to the Forth type (e.g., off_t would
always be a double, whether the basic type is long or long long an the
actual platform).  Moreover, the idea is that someone would use that
tool to generate the Forth declarations for a given API once (and
possibly apply manual corrections), and from then on we would use the
file containing the Forth declarations whenever we want to interface
with that API.

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: http://www.forth200x.org/forth200x.html
   EuroForth 2008:
http://www.complang.tuwien.ac.at/anton/euroforth/ef08.html