Path: utzoo!utgpu!watserv1!watmath!att!pacbell.com!mips!sgi!duncan@spectre.vlsi.sgi.com
From: duncan@spectre.vlsi.sgi.com (Donald Duncan)
Newsgroups: comp.os.minix
Subject: Re: GCC and GNUEMACS for MINIX-386
Summary: Fantastic, Great, Thank You
Message-ID: <76457@sgi.sgi.com>
Date: 29 Nov 90 19:37:46 GMT
References: <3469@aipna.ed.ac.uk>
Sender: guest@sgi.sgi.com
Organization: Silicon Graphics, Inc., Mountain View, CA
Lines: 53

In article <3469@aipna.ed.ac.uk>, awb@aipna.ed.ac.uk (Alan W Black) writes:
> 
>          Announcing gcc 1.37.1 and Gnu emacs 18.55 
>                   for MINIX 1.5.10 386
>                     for beta testing
>...

Thank you very much!  Merry Christmas to you too!  This was one of the last
missing pieces of the puzzle.

>...
> It is interesting to note that gcc and gas effectively require
> no patches at all to compile under MINIX.  Gas requires a change to 
> the make file while the gcc diffs simply include a modified config file
> for the 386.  Our problem was bootstraping the first copy of gcc.

No kidding.  I pulled most of my remaining hair out trying to get gcc
bootstrapped.  I set it aside about two months ago when classes resumed.

I installed the bin, lib, and include files (late) last night.  So far I
have found one error in your instructions.  In gcclib/README you say:

>	    Libraries for gcc minix implementation
>
>	    These libraries should go in the following places
>
>	       cp libc.a /usr/lib/libc.a
>	       cp libm.a /usr/lib/libm.a
>	       cp gnulib /usr/local/lib/gcc/gnulib
>	       cp crt0.o /usr/local/lib/gcc/gnulib


The last line should read:


>	       cp crt0.o /usr/local/lib/gcc/crt0.o

The only substantial things I have compiled so far are bison and flex.
Flex had minor problems in the prototype declaration for unlink (trivial
fix) and bison compiled and loaded on the first attempt.  I did have to
modify the Makefiles to search /usr/local/lib/gcc/gcc-include before
/usr/include.  This may be due to the particular installation state that
I have - I'm not certain yet.

I'm having difficulties with the floating point library but I'm sure it
is a case of operator error.  If it isn't I'll post bugs/fixes.

Again, Many Thanks to all involved.




duncan@sgi.com (Donald Duncan)