🚀 go-pugleaf

	Yes, getting a port of 386bsd-0.1 would be a great thing to do.  I have
	actually played with 386bsd (on a pc for 1 week) and I have the sources

What is this in reference to?  I seem to have missed the thread.
Gary and I were talking about this just a couple of weeks ago.

I also run 386BSD on my PC (I have for the last month and a half or so and
I'm quite happy with it).

I've had gcc producing reasonable bsd binaries for the '532 for some
time (it's not hard).  When linked with my stand-alone stdio library,
they actually run, too.  None of this is much of a trick, however.
The hard part (as I see it) is extracting the MMU setup code from the
mach port and somehow grafting it into 386BSD, as well as autoconf'ing
in the serial and scsi drivers.  Most if the 32532 specific math and
assembly code for libc can probably be stolen outright (in fact, the
386BSD math libraries compile with -DNATIONAL!).

I need to look at how much trouble the drivers are going to be, but
this is hardly my area of expertise.  Gary would be far more qualified
to comment on that.

Actually, I think this would probably amuse the hell out of Bill
Jolitz, and he could possibly even be halfway helpful with it if we
asked him nice.  Those familar with Bill's history will know why it
would amuse him to see 386BSD on a national chip.

One ofh the problems I see is hosting all this in a coordinated
fashion.  Phil appears to be trying to host from minix whereas I'm
hosting from a PC that shares the same SCSI disk (makes populating a
BSD fs with cross-compiled binaries a lot easier!).

Adding 30 character filenames to minix is very easy - I'm surprised
Phil hasn't done that particular hack yet.  In any case, I think
cross-compilation is the way to go, for now.

					Jordan