Re: Ada for the ARM?

[Steffen Huber <hubersn@lcs.wn.bawue.de>]

In message <7eo48a$1ocm$1@odie.mcleod.net>
          RSWhite@nospam.somewhere.ia.us (Robert S. White) wrote:

>   I'm curious to know if any of the Ada compiler vendors are working
> on an Ada compiler for the Advanced Risc Machines ARM processor.

I am curious too...

> With the upcoming advent of the ARM10 which has a vectored floating
> point math coprocessor option, it would seem to be a natural for
> the larger embedded math heavy applications that have minimal power
> consumption and cooling requirements.  There is a gcc port for the
> earlier generation ARMs so a GNAT variant should be readily do-able.

Actually, there is already a GNAT port to RISC OS, ARMLinux and
NetBSD/arm. All of these were done by Peter Burwood. You could try
to contact him at gnat@arcangel.dircon.co.uk - be aware however
that Pete is currently suffering from some type of RSI, and is only
now and then able to read (and answer) his email.

The ports were based on GNAT 3.03 (at least the RISC OS one - Pete had
3.09 nearly ready when I last contacted him).

The RISC OS variant that I am currently using for CDBurn development
uses 26bit ARM code of course, which is useless for the ARM10, as the 26bit
processor modes got scrapped. And I guess the ARM backend of gcc would need
to be updated to support the math coprocessor - IIRC the opcodes are
different from the earlier ARM FPU opcodes (I might be wrong here). Richard
Earnshaw from ARM is the maintainer of the gcc ARM backend, so he would be
the man to talk to.

> Till then I guess the Avenstar (Intermetrics) C to Ada compiler is
> the only option.

This is always an option of course.

So long, Steffen

-- 
Steffen Huber           LambdaComm System - Welcome to Trollinger Country
hubersn@lcs.wn.bawue.de                       acorn2@ftp.uni-stuttgart.de
... Mehr Recht fuer den Hecht, mehr Rasen fuer den Hasen.