Path: utzoo!attcan!uunet!mcvax!botter!star.cs.vu.nl!ast
From: ast@cs.vu.nl (Andy Tanenbaum)
Newsgroups: comp.os.minix
Subject: Re: Status of V1.3
Message-ID: <719@ast.cs.vu.nl>
Date: 13 May 88 10:36:57 GMT
References: <709@ast.cs.vu.nl> <4530@hoptoad.uucp>
Reply-To: ast@cs.vu.nl (Andy Tanenbaum)
Organization: VU Informatica, Amsterdam
Lines: 90

In article <4530@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes:
>Does the anticipated Minix V1.3 have working serial ports?
At this moment, I know of 5 or 6 people who are working on serial ports.
None of them work yet.  Some crash, some need obscure tuning, some are not
plug compatible with the old tty.  The latter is an important issue due to
many changes made in the kernel both as a result of the Ethernet stuff which
has been added and the Atari stuff.  I don't have the time to write a 
serial driver myself right now, but if someone else can supply me with a
FULLY DEBUGGED tty.c file supporting RS232, I am all for it.


>Since it has compress and uu**code, it should have atob/btoa too.
I don't have them.  Somebody please send/post them, along with documentation.

>The compress documentation should say that it's a 12-bit compress (if it is),
>not the standard 16-bit compress.
Beats me.  It uses Lempel-Ziv coding, whatever that is.  I got it from the net
and it seems to work.  Average file is cut in half when compressed.

>A much better file(1) command is in the comp.sources.unix archives,
Its 110K.  Mine is 3K.  The Berkeley spirit at work: 3 times better and 30
times larger.  Bad tradeoff.  It is this mentality that makes my 512K MINIX AT 
at home apleasure to  with, whereas I curse a blue streak at my SUN-3 at work
work because it has only 4M on it.

>The super fast James Woods grep/fgrep/egrep from comp.sources.unix should be
>supplied, too.
Could some kind soul take a look at this, see how big it is, and if it is not
absurd, port it to MINIX and post it. If it don't fit in 64K, it don't count.

>Patch(1) by Larry Wall should be supplied rather than the "fix"
If somebody will do the work and provide me with a fully debugged version
for MINIX and it is ok with Larry, I'll include it in addition to fix.

>A real version of tar(1) should be included, e.g. one that uses the same
>arguments as v7 tar.  Mine is in comp.sources.unix.
I tried one other tar; I forget whose, but it had a lot of problems.  If
yours has been thoroughly tested on MINIX, please send it to me to.  Again, it 
must fit in 64K.  Don't talk to me, talk to Intel.

>Has anyone retrofitted the Berkeley TCP/IP networking code recently posted? :-)
>Or even Phil Karn's stuff, which already runs on MSDOS IBM PC's?
"Not I," said the little red hen.

>I might suggest that just reading the index to comp.sources.unix and
>obtaining anything that looks like a PD reimplementation of a Unix command
>is likely to produce a much more functional Minix.  The c.s.u code was
>written by and for people on real Unix systems, and was edited by good
>moderators, so it is usually as useful as real Unix code, unlike
>much of Minix.
My experience is different.  Every time I try to take something from that
newsgroup, I end up throwing it away because it either has a makefile a
yard long dealing with 4.1, 4.2, 4.3, System 5.0, System 5.1, System 5.3,
V6, V7, and various other systems, it uses all manner of strange library
routines I have never heard of (no doubt 4.something or 5.something) or
it barely fits on my 4M SUN with its lousy little 70M disk, let alone my
impoverished AT with its minuscule 20M disk, or something else is wrong.
I have gotten many useful programs from readers of this group (thanks!) but
very little from comp.sources.unix.
Remember, the goal of MINIX is to have a system that is small enough to
understand.  If you just want to run UNIX, buy Xenix.


>Bring the standard library up to snuff so that it can compile the
>average 1988 program posted to the net, e.g. all the string functions,
The /usr/include directory has been reorganized to be more like true V7.
As far as new library routines, if somebody wants to fish them out of
archives, test them thoroughly on MINIX, and send them to me, fine.
Again, my experience is that everytime I start on something like this, I end
up discovering that nothing works because all the routines make 4.3 or 5.3
calls.

>Upgrade the C Preprocessor to handle #if defined() and other things.
>ANSI C would be best of course.  There are two working ANSI C
>Preprocessors available free in source (gnu and decus).
The compiler works quite well and is internally consistent.  I am not
inclined to change it now.  As to GNU being free,  we have gone through
that issue before at great length.  Perhaps Stallman and the combined legal
departments of Prentice-Hall, its parent company Simon & Schuster, and ITS
parent company, Gulf+Western could work that out :-) .

>Porting my tar to Minix involved writing execlp(), bcmp, bzer(),
>3 arg open.
That is exactly my point.  To use some utility, one must first implement the
4.3 library, and then the 4.3 operating system.  Next time I give my operating
systems course, I'll check to see if any students want to implement 4.3 for
extra credit.  Last time I asked if any wanted to implement MS-DOS.  There
were no takers.  On that note, I will depart.

Andy Tanenbaum (ast@cs.vu.nl)