Path: utzoo!attcan!uunet!nuchat!abbadon
From: abbadon@nuchat.UUCP (David Neal)
Newsgroups: comp.sys.amiga
Subject: Re: Ghostscript
Summary: Stallman
Message-ID: <2654@nuchat.UUCP>
Date: 9 Nov 88 06:51:45 GMT
References: <9523@watdragon.waterloo.edu> <0XRmvTy00Uka0TS645@andrew.cmu.edu>
Distribution: na
Organization: Public Access - Houston, Tx
Lines: 36

In article <0XRmvTy00Uka0TS645@andrew.cmu.edu>, bader+@andrew.cmu.edu (Miles Bader) writes:
> bmacintyre@watsol.waterloo.edu (Blair MacIntyre) writes:
> > Also, anyone working on GCC or GC++??
> 
> Of course, to run ANY program written by Richard Stallman, you'll
> need an awful lot of contiguous memory so that the "stack 1000000"
> command will work...


Absolutely true... this is what convinced me porting gnu cc and 
gnu cc++ would be a good way to waste a major fraction of 
my lifespan. I hear GNU has a kernel more or less working 
(or at least compilable) and it takes many many megs of space
to do so. Now of course, so does most kernels these days,
SUNOS takes plenty of space, but considering that someone
has posted a UNIX (tm) clone that runs in  ->32K<- and can run a fair
amount of unix software it would follow that with a resonable
effort, gcc and gcc++ could be considerably smaller.
 
Beyond this, gcc still depends on YOUR libraries in YOUR
compiler, so be prepared to write your own if you wish
to distribute an entire compiler that's public domain.

Anyone still up for
           1) converting gas to spit out proper amiga .o files?
           2) converting gld to understand scatter loading and
              hunks?
           3) writing your w own crt.c libraries? Adding all the
              'c' library routines? (remember, this thing has
               to understand launching a task)
           4) Buying 4 megs to compile your 'C' compiler in?

If so, mail me, any excuse to buy more memory is ok by me! :-)

David Neal
killer!nuchat!abbadon