Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!rice!sun-spots-request
From: charyb!dan@uunet.uu.net (Dan Mick)
Newsgroups: comp.sys.sun
Subject: Re: ld.so dirties user stack
Keywords: Miscellaneous
Message-ID: <5066@brazos.Rice.edu>
Date: 13 Feb 90 17:51:16 GMT
Sender: root@rice.edu
Organization: Sun-Spots
Lines: 14
Approved: Sun-Spots@rice.edu
X-Refs:  Original: v9n36, Replies: v9n40
X-Sun-Spots-Digest: Volume 9, Issue 50, message 1

In article <4892@brazos.Rice.edu> gandalf@csli.stanford.edu (Juergen Wagner) writes:
>X-Sun-Spots-Digest: Volume 9, Issue 40, message 1
>
>Hmmm... Don't rely on such things as local variables being
>zero-initialized.  Otherwise, you'll wind up with mysterious errors on
>some machines.

I don't know of *any* machine that assures that local (automatic, stack)
variables are 0!  You may get lucky on VM machines if that's the first
time you've called down that far on the stack, but it's just chance, and
it sure won't be a "benevolent side-effect" on anything like a Sun-4,
which puts most or all of its automatics in registers...

Shudder!  Autos *must* be initialized.  Always.