Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!think!ima!haddock!karl
From: karl@haddock
Newsgroups: net.lang.c
Subject: Re: Significant deficiency in C
Message-ID: <86900071@haddock>
Date: Thu, 2-Oct-86 10:33:00 EDT
Article-I.D.: haddock.86900071
Posted: Thu Oct  2 10:33:00 1986
Date-Received: Sat, 4-Oct-86 06:40:37 EDT
References: <4200@brl-smoke.ARPA>
Lines: 19
Nf-ID: #R:brl-smoke.ARPA:4200:haddock:86900071:000:909
Nf-From: haddock!karl    Oct  2 10:33:00 1986


00R0DHESI%bsu.cs@CSNET-RELAY.ARPA (Rahul Dhesi) writes:
>I think [char to int promotion] is a significant deficiency in C.  Is ANSI
>doing anything to eliminate this or decrease its effect?

I doubt it.  Too many programs depend on it.

>Considering that [most] UNIX tools do text processing in 7-bit units, ...

But if they use getchar(), they have to declare the variable "int" anyway.

However, a compiler is free to use 8-bit arithmetic if it can determine that
the result will be the same (e.g. char2 = char0 + char1); I think more work
needs to be done in this area.  (Of course, there will always be cases where
it lacks the information to perform the optimization.)

Karl W. Z. Heuer (ima!haddock!karl or karl@haddock.isc.com), The Walking Lint
P.S. I just tried c2=c0+c1 on the vax; it did cvtbl, cvtbl, addl2, cvtbl, and
cvtlb.  This sort of slop encourages programmers to declare everything "int".