Path: chuka.playstation.co.uk!news
From: "pal" <pal_news@hotmail.com>
Newsgroups: scee.yaroze.freetalk.english
Subject: Re: Ten bytes bug in gcc?
Date: 14 Aug 2001 21:58:14 GMT
Organization: PlayStation Net Yaroze (SCEE)
Lines: 21
Message-ID: <01c1250b$c7e92d60$501ce4d5@pal-s-omnibook>
References: <01c124b1$16cc8cc0$LocalHost@pal-s-omnibook> <9lbptg$fio2@www.netyaroze-europe.com>
NNTP-Posting-Host: nas-cbv-2-28-80.dial.proxad.net
X-Newsreader: Microsoft Internet News 4.70.1155

Omar <OMetwally7@aol.com> wrote in article
<9lbptg$fio2@www.netyaroze-europe.com>...
><snip>you might want to make it unsigned char test[11] instead so that
> the last one is NULL value.
><snip>

Standard libraries string functions (such as printf, scanf, strlen...)
require and produce strings of chars ended by '\0'; and string constants
end with '\0'. This is the reason why strings usually end with an '\0'. But
here the array is not meant as a string, it's just a piece of memory.

The type char/unsigned char is always 8bit wide, this is why I use it in
the example, so as to be sure of the physical size of the array. Originally
I encountered the problem with a ten-byte struct.

In the faulty code no use is made of the array, the declaration suffices to
cause the crash, this is the weird thing.


pal