On Wed 10 Dec 1997, maximum entropy wrote:
> >From: hyc@troy.la.platsol.com (Howard Chu)
> >
> >  I think it would be nice if I could compile the TOS utilities
> >  (loadbsd.ttp, aptck.ttp, etc) for NetBSD/atari under NetBSD, instead
> >  of requiring a cross-compilation setup or compilation under MiNT.

This is handy sometimes...yes.

[ ... ]

> >  Actually getting real programs to compile in this manner would be a
> >  lot of work.  Does anyone else think this is a good idea?  If so, i'll
> >  persue it further.  I'd hate to do all the work only to find out that
> >  it would never be integrated into the tree.

If you manage to get it done, I'll happily integrate it in the tree. I never
felt the urge of 'fixing' this, but if you do, feel free ;-)

[ ... ]

> The goal I have in mind here is to provide a system where a user would
> not need to collect together little bits and pieces from here and
> there to compile a fully functional NetBSD/atari system from source.
> As it is one needs to set up cross-compilation (or at least
> cross-linking), get the MiNT libraries if one doesn't already have
> them, collect the XHDI source from somewhere and turn it into a
> library, install gmake if you don't already have it, etc etc.  In my
> opinion this is very unclean compared to the NetBSD/i386 system, where

This is very much true. The above is the sole reason I keep my old MiNT
compile environment. Since I was able to run some kind of UN*X system on
my TT, MiNT is only used to compile 'tostool'-like things (once or twice a
year).

> I can simply type "cd /sys/arch/i386/stand && make cleandir && make
> depend && make && make install" and out pops a new copy of dosboot.com
> in /usr/mdec along with the bios bootblocks.  Similarly, I don't need
> to set up an ULTRIX cross-compilation or cross-linking environment to
> build an a.out or ECOFF standalone kernel under NetBSD/pmax that can
> be booted from the pmax boot PROMs.
> 
> I see no reason not to support the same type of thing in the
> NetBSD/atari tree, except that it will take some work to get it right.
> Most of the work would involve figuring out some minimal subset of the
> MiNTlibs and headers, and XHDI code, that we would need in the tree to
> support this.  Since all the required code is public domain, this
> should not be problematic with respect to distribution restrictions.

Except that most of it bears an FSF copyright I think...

Leo.