From owner-freebsd-hackers  Sat Apr 15 12:11:27 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id MAA21187
          for hackers-outgoing; Sat, 15 Apr 1995 12:11:27 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id MAA21180
          for <freebsd-hackers@FreeBSD.org>; Sat, 15 Apr 1995 12:11:18 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id EAA32482; Sun, 16 Apr 1995 04:53:56 +1000
Date: Sun, 16 Apr 1995 04:53:56 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199504151853.EAA32482@godzilla.zeta.org.au>
To: freebsd-hackers@FreeBSD.org, j@uriah.heep.sax.de
Subject: Re: 90's compilers
Sender: hackers-owner@FreeBSD.org
Precedence: bulk

>> If you're proposing changing the coding guidelines away from KNF
>> on this I'm all for it.

>There has never been an `official' coding guideline to keep KNF.  In
>fact, some device drivers under i386/isa do already fully use `new'
>style.

These drivers demonstrate that there is no `official' coding style,
not that there shouldn't be one.

>The boot code has been crying for a revamp.  The writers didn't even
>bother to declare external functions at all (``all the world is an
>int, and might take any possible number of arguments'' :).  There have
>been many errors in the code (functions that actually take a `char'
>argument, while the compiler promoted it to `int' -- works only since
>we're little-endian).

>This style made it highly error-prone.

Actually the functions took an `int' arg and demoted it to a `char'.
This is perfectly valid K&R1 and ANSI code.  Declaring the functions
as taking a `char' arg changes the interface are requires prototypes
to be in scope when the functions are called.

Bruce