Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!cbatt!ihnp4!inuxc!iuvax!ndmath!milo
From: milo@ndmath.UUCP
Newsgroups: comp.unix.xenix
Subject: Re: lex problems on Altos
Message-ID: <190@ndmath.UUCP>
Date: Sat, 21-Feb-87 18:02:12 EST
Article-I.D.: ndmath.190
Posted: Sat Feb 21 18:02:12 1987
Date-Received: Mon, 23-Feb-87 01:48:09 EST
References: <59100002@gorgo.UUCP>
Organization: Math. Dept., Univ. of Notre Dame
Lines: 20

In my experience, Altos Xenix stinks really bad, I did a review on it for
Byte awhile back...but lately I have been noticing even more problems

Recently I tried to compile EMPIRE on the Altos, it ran but has many unexplanable
bugs (which don't appear on a VMS compiled version).  I spent a great deal of 
time hunting down one of the bugs and finally found that the bug was actually
in the C compiler...when I removed references to register variables the code
ran fine.

I have also had the Altos Xenix compiler barf on many other public domain programs
including Super Rogue (preprocessor died), news 2.10 (compiled but running postnews
caused system crashes) and several others.  All the programs mentioned above ran
fine on both a SUN and a VAX.  We also tried compiling some other programs on
SCO Xenix V on an AT and had similar problems.

Personally, I think the Xenix C compiler stinks!  Commercial applications
seem to work more-or-less ok but if you want to run any public domain stuff
you might as well forget it.

Greg Corson