Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!rochester!ritcv!ccivax!rb
From: rb@ccivax.UUCP (rex ballard)
Newsgroups: net.lang.c
Subject: Re: casting structures
Message-ID: <430@ccivax.UUCP>
Date: Mon, 24-Feb-86 13:46:29 EST
Article-I.D.: ccivax.430
Posted: Mon Feb 24 13:46:29 1986
Date-Received: Sat, 1-Mar-86 01:08:50 EST
References: <302@hsi.UUCP> <1450@bbncc5.UUCP> <305@cray.UUCP> <269@spp3.UUCP> <392@ccivax.UUCP> <165@dg_rtp.UUCP>
Reply-To: rb@ccivax.UUCP (What's in a name ?)
Organization: CCI Telephony Systems Group,  Rochester NY
Lines: 42

Well, another goofup!
First of all, the example given was poor to say the least.  It certainly
did not illustrate the problem.  Your right, I've never used it, but I
have seen it.

An example where casting was needed was this.
#define EOR 0370
char *x;
if (*x<EOR)
or
#define EOR '\370'
int *x;
if (*x<EOR)

(believe it or not, because of some #includes, both showed up in the same
project, in fact, the definitions changed mid-project)

guess what, the original compiler actually worked (gave the result desired)
with the second example and didn't work with:

#define EOR '\370'	/* compiler treated like unsigned */
char *x;
if(*x<EOR)

because of the nature of what was actually desired, the proper definition
would have been.

#define EOR (unsigned char)0370
or
#define EOR (unsigned char)'370'
unsigned char x;
if (*x<EOR)

moral: when in doubt, cast  (the next guy might figure out what you want)

Qualifier: the compiler that gave this problem accepted code which
contained so many syntax errors (missing semicolons...) that lint
died just looking at it.  (I think it was someone's kludge of bcpl or
what they thought C was before reading K&R :-)  It is a VERY OLD compiler.

Please excuse problems in the code, these are just "off the cuff" examples,
written near the end of the lunch hour.