Subject: Re: Cross-developing MacBSD
To: None <>
From: Brad 'Savory Caramel Coating' Grantham <>
List: macbsd-general
Date: 12/02/1993 01:41:07
(Sorry, John; you got another copy of this accidentally, but I corrected
your name for this one.)

John Kuzma writes:
> It seems to me that one should be able to take the MacBSD code, load it up
> under something like Linux or NetBSD on a PC, create a version of gcc as a
> cross-compiler, and work on developing support for the as yet unsupported
> Macs. Transfering the kernel image could be done serially or via floppies and
> something like Suntar. My point in all of this is that there are a lot of 
> different Macs out there, and apparently, very few developers at this point.
> Perhaps a few of the would-be developers like myself can help out in the porting
> effort now without having a working kernel for their own machine (like my IIsi).

Yes, this definitely possible, especially under NetBSD/386.  In fact, I've
considered doing the same thing on my 486/66, since my 16MHz Mac II
crawls through NetBSD/Mac builds.  I guess I want to say two things about

1) The primary developers are already having a bitch
of a time getting just the basic utilities compiled, on the native
system!  I think that, though it is possible to develop a cross-
development system under, say, Linux for NetBSD/Mac, it may be kind
of pointless.  It requires porting GCC, GAS, and the GNU binutils package
to your host OS, managing the MacBSD include and source tree, and
understanding the kernel code without a working model.  Finally, you
can hardly tell the difference between a kernel that doesn't work and a
compiler that generates incorrect output.  It sounds to me like we'll
be done with the port for your machine around the time you get your
cross-compiler working.  (This should inflame a lot of people into

2) Having said that, I encourage anyone willing to try to go ahead.
When we started out with A/UX 2.0 and a tape of Net/2 almost two
years ago, I knew next to nothing about free software, portability
issues, or UN*X in any depth.  (And now I know twice that much! :)
Trying to port even just GCC 1.39 to an as-yet nonexistent BSD platform
was an eye-opener and an education.  I will remember the summer of
92 and the following school year with fondness for the rest of my
life.  If you have the resources (disk space, time, a little
inventiveness), by all means you should try.  (And I mean any large
(fun) project, not just NetBSD/Mac.)


(I apologize for the inordinately high noise-to-signal ratio of this