4.2 minix and request for news group

[Dan Lanciani]

In article <1744 at mulga.OZ>, andrew at mulga.OZ (Andrew Worsley) writes:
>     As you probably know the main (and large cost) for non educational
> institutions buying 4.2 (source and binary) is buying the required AT&T
> source licenses, which is neseccary as 4.2 contains V7 (I believe it is
> V7) source code. How hard would it be to build a 4.2 based on minix's
> V7 code taking the 4.2 code from Berkeley's work?  I am sure that
> Berkeley would be happy to be released from their bonds to AT&T. How
> easily can one do this legally, I mean how can you seperate out code
> written by Berkeley and that by AT&Ts? Can you legally work from code
> that is AT&Ts and publish it?

	As I posted earlier, I have already written a non-AT&T kernel that
supports virtually all 4.3 system calls (sans network).  This includes
all the obscure tty ioctls and modes, job control, signals, and even the
a.out header.  The notion of image memory is sufficiently general that
a port to the vax should not be impossibly difficult.  Support for
generic file systems is there, so dropping in NFS would be fairly easy.
Getting a descent tcp implementation might be a problem, but I will
be adding udp/ip as time permits.  I am most interested in getting it
up on a 386.
	This kernel is now running on my Z100 and a Tandy PC clone.  The
PC version uses the ROM BIOS (mostly) and should run on work-alikes as
well.  When I am convinced the code is fairly solid, I will post a
note in case there is interest.

				Dan Lanciani
				ddl at harvard.*