Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!asuvax!ncar!ico!haddock!karl
From: karl@haddock.ima.isc.com (Karl Heuer)
Newsgroups: comp.std.c
Subject: Re: memcpy
Keywords: memcpy
Message-ID: <18083@haddock.ima.isc.com>
Date: 20 Sep 90 17:55:56 GMT
References: <1990Sep19.021418.11574@maths.tcd.ie>
Reply-To: karl@haddock.ima.isc.com (Karl Heuer)
Organization: Interactive Systems, Cambridge, MA 02138-5302
Lines: 17

In article <1990Sep19.021418.11574@maths.tcd.ie> tim@maths.tcd.ie (Timothy Murphy) writes:
>In [unix] unzip.c it is assumed that the effect of
>	buf[0] = c;
>	memcpy(buf+1, buf, 20);
>is to set
>	buf[0] = buf[1] = buf[2] = ... = buf[21] = c.

This is incorrect.  The proper way to achieve that effect is
	memset(buf, c, 21);

>Is this a bug on the Sparc [which doesn't do this],
>or is memcpy not fully specified?

The latter.  The effect of memcpy() on overlapping arrays is undefined.
(There is a new function memmove() which does have predictable behavior.)

Karl W. Z. Heuer (karl@kelp.ima.isc.com or ima!kelp!karl), The Walking Lint