Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!netnews.upenn.edu!cs.widener.edu!news.cs.indiana.edu!samsung!caen!hellgate.utah.edu!dog.ee.lbl.gov!elf.ee.lbl.gov!torek
From: torek@elf.ee.lbl.gov (Chris Torek)
Newsgroups: comp.lang.c
Subject: Re: Question about pointer
Message-ID: <10497@dog.ee.lbl.gov>
Date: 2 Mar 91 00:45:36 GMT
References: <2488@njitgw.njit.edu>
Reply-To: torek@elf.ee.lbl.gov (Chris Torek)
Distribution: usa
Organization: Lawrence Berkeley Laboratory, Berkeley
Lines: 27
X-Local-Date: Fri, 1 Mar 91 16:45:36 PST

In article <2488@njitgw.njit.edu> yxz1684@hertz.njit.edu (Ying Zhu ee) writes:
>The following is a small program from "Mastering C pointers" by Robert J.
>Traister. [example deleted] ... Since the r[100] is an automatic variable,
>after calling the "combine" the r[100] will be deallocated [and] p will
>point to an unsafe place.

Correct.

If the example in this book appears exactly as shown in
<2488@njitgw.njit.edu>, I would suggest either discarding the book
entirely (reselling it merely pushes the bogus information on someone
else) or---if this is atypical, and merely one mistake out of a great
deal of correct code---advising the author and/or publisher, and
marking the example in the book to note its error or repair it.
(One `static' will repair this particular sample, for instance.)

Note that the only excuse these days for buggy examples is when the
publisher do their own re-typesetting, since book examples should
be typeset by direct insertion of working code into the typesetting
system.

There are many C books available these days, and a large fraction of
them are inexcusably bad.  Do not trust any single source (except
perhaps a copy of the ANSI C standard X3.159-1989).
-- 
In-Real-Life: Chris Torek, Lawrence Berkeley Lab EE div (+1 415 486 5427)
Berkeley, CA		Domain:	torek@ee.lbl.gov