Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!ucbvax!ANDREW.CMU.EDU!gk5g+
From: gk5g+@ANDREW.CMU.EDU (Gary Keim)
Newsgroups: comp.soft-sys.andrew
Subject: Re: Unaligned access on DECstation
Message-ID: <IbvZIrS00VsWEL4J8H@andrew.cmu.edu>
Date: 25 Mar 91 19:53:27 GMT
References: <IbvYs59O_U00E4TGUP@chillon>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 24

Excerpts from misc: 25-Mar-91 Unaligned access on DECstation Jeff
Putsch@chillon.uicc (1064)

> I don't have gdb, so I can't debug ATK and track the problem. Does
> anyone know what needs fixing?


There is someone here in Academic Computing who is investigating this
error and how to make it go away.  You see, CMU bought a whole bunch of
pmaxen and then found out that Andrew didn't really perform on that
platform.  The shell would often spit out error messages when running an
ATK program.  It would also core dump at random times.  It turned out to
be a memory management bug in the kernal and a bug with the exec()
system call.

This "unaligned access" message apparently results from bad programming
practice.  I have a pmax in my office and I've never seen this message. 
Beats me.

When we make the fixes that cause this message to go away, we'll report
on the cause.

Gary Keim
ATK Group