Re: [MiNT] MiNTLib for ColdFire

[Lonny Pursell <atari@bright.net>]

on 4/29/09 5:47 AM, Vincent Rivière wrote:

> There is not so much work to do in a compiler/interpreter.
> Some examples :
> 
> - replace addq.w by addq.l
> 
> - replace move.l d0-d7,-(sp)
> by  lea -32(sp),sp
> move.l d0-d7,(sp)
> 
> and avoid missing instructions like rol, etc...
> 
> For operating systems, it is a little more tricky because of different
> stack frame handling on traps, but for normal apps there is no such problem.
> 
> So sure, if you add native ColdFire support in the GFA it would be a
> definitive good feature for ACP users !

Interesting. Thanks for the examples.  The reason I think it's a huge task,
if because the entire sources are 100% assembler.  I'll rephrase it as
perhaps tedious work then.  ;-))

As for the discussions about other paths, it's indeed a subjective thing.
Clearly everyone has their own idea of what will satisfy them personally and
the old phrase comes into play... You can't please everyone.  I think that
is very much the case here no matter which way the new hardware would go.

I don't see CF as a totally new CPU, and it's probably how many see it, you
only have to partially fix/port your code, it's not a 100% rewrite and you
stick with a language you already know.  Assuming one is into 68k
programming.  

I also see this type of discussion on all forums, ditch the 68k, port it to
ppc, intel, arm, whatever processor is currently the in thing.  It's all
well and good, but the bottom line is no one ever actually does it. They
only suggest it.  Now if someone actually done it, then dropped it on the
community, it might well take off and even win over folks against it.
Especially if it was able to run old programs seamlessly via a robust
emulation layer.

Bottom line is simple for me at the end of the day.  I am unable to build my
own clone, so I just have to decide if I'm going to support whatever appears
next.  To that I say, yes. :o)

-- 
Lonny Pursell    http://www.bright.net/~gfabasic/