A/UX Release 2.0

[John Coolidge]

liam at cs.qmw.ac.uk (William Roberts) writes:
>gday at digigw.lab.digital.co.jp (Gordon Day) writes:
>A/UX 2.0 is still System V at heart, in that it uses inittab
>rather than just a single /etc/rc script, you have to think
>about using cpio (though tar is available), rsh means
>"restricted shell" and not "remote shell". Apart from that it
>isn't any more of a nightmare than moving from SunOS to other
>non-Sun machines.

Actually, it's been much easier than a number of other non SunOS machines
I could name. We've got a very heterogeneous lab, and the A/UX machines
have been about the easiest to integrate with the Suns (which tend to be
the reference standard). Administration-wise, they're a bit different
(inittab rather than /etc/rc, etc), but not much of a problem.

>>Is the X option available from Apple really just the standard release with
>>yet another TWM lookalike?

>YOu could compile the standard release if you really wanted to
>be sure of what you were getting. MacX is the thing where Apple
>have done some work on producing a psuedo-rooted system (i.e.
>the root window is not done with X, and the window manager is
>provided by the native Mac system).

To expand on this a bit: Apple has two X products. X11R{3,4} for A/UX is,
basically, just the MIT X product. R3 adds support for color to MIT's
R3, but since MIT's R4 is faster than R3 and has color support, there's
not much reason to run R3. We're running R4 and doing just fine.

MacX, on the other hand, runs under MacOS or the A/UX Finder, and provides
a Mac-like interface to X. The normal Mac interface is mapped into a ICCCM
compliant window manager, and X windows can be either rooted (in a normal
Mac window which functions as a X root window) or rootless, in which case
they operate very much like normal Mac windows.

>>How buggy is the current release, and with the new features, what's the guess
>>on the next release?

>A/UX 1.1.1 is solid though its Mac application support is
>fairly limited. We've used it for several months with a student
>lab of 100 machines and had very few panics. Some of the
>utilities aren't so hot and the C compiler isn't nearly as good
>as gcc. As for guesses about A/UX 2.0, guess away...

The normal 'cc' that comes with A/UX is pretty good iff you only want
to compile small to medium programs. It has a few minor problems with
some dubious C syntax, but by and large it's pretty solid. The biggest
problem is has are fixed-length tables (for symbols, etc), which simply
overflow on large programs. However, gcc 1.37 is very solid on A/UX (I've
compiled X11R4 and other large programs with it and had no problems).
You still have to live with Apple's assembler, though :-( (it produces OK
output, but has fixed size tables too and sometimes loses optimizations,
and has a _very_ limited symbol space). Gas doesn't provide support for
COFF yet...

>>I plan on running a number of IIci's with X on A/UX together with Sun 4/80s so
>>I would like to mount the A/UX /usr on a Sun server.  Will the A/UX side be
>>happy with this?

>No problems at all - we did almost exactly this (but with a VAX
>11/750) for some while before putting in a IIcx as the server.

We've got a similar configuration, with /usr/local, /home (our home
directories), and optional stuff mounted on foreign machines (all
kinds of different machines --- NFS works just fine). The only things
that _have_ to be local are swap space and system programs (A/UX
can't boot diskless).

--John

--------------------------------------------------------------------------
John L. Coolidge     Internet:coolidge at cs.uiuc.edu   UUCP:uiucdcs!coolidge
Of course I don't speak for the U of I (or anyone else except myself)
Copyright 1990 John L. Coolidge. Copying allowed if (and only if) attributed.
You may redistribute this article if and only if your recipients may as well.