From owner-freebsd-bugs  Sat Apr 18 07:39:56 1998
Return-Path: <owner-freebsd-bugs@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id HAA18749
          for freebsd-bugs-outgoing; Sat, 18 Apr 1998 07:39:56 -0700 (PDT)
          (envelope-from owner-freebsd-bugs@FreeBSD.ORG)
Received: from elvis.vnet.net (elvis.vnet.net [166.82.1.5])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA18740
          for <freebsd-bugs@hub.freebsd.org>; Sat, 18 Apr 1998 14:39:53 GMT
          (envelope-from rivers@dignus.com)
Received: from dignus.com (ponds.vnet.net [166.82.177.48])
          by elvis.vnet.net (8.8.8/8.8.4) with ESMTP
	  id KAA02623; Sat, 18 Apr 1998 10:39:38 -0400 (EDT)
Received: from lakes.dignus.com (lakes [10.0.0.3])
	by dignus.com (8.8.5/8.8.5) with ESMTP id LAA25318;
	Sat, 18 Apr 1998 11:07:51 -0400 (EDT)
Received: (from rivers@localhost) by lakes.dignus.com (8.8.7/8.6.9) id KAA09701; Sat, 18 Apr 1998 10:44:21 -0400 (EDT)
Date: Sat, 18 Apr 1998 10:44:21 -0400 (EDT)
From: Thomas David Rivers <rivers@dignus.com>
Message-Id: <199804181444.KAA09701@lakes.dignus.com>
To: jkh@time.cdrom.com, rotel@indigo.ie
Subject: Re: bin/6332: time.h doesn't compile with -traditional
Cc: bde@zeta.org.au, freebsd-bugs@hub.freebsd.org
Sender: owner-freebsd-bugs@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

> 
> > Couldn't agree more, all new code should use
> > 
> > 	int main(int argc, char** argv)
> 
> Or even:
> 
> int
> main(int argc, char *argv[])
> 
> If you ever want to create tags for the code. ;)
> 
> 				Jordan
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-bugs" in the body of the message
> 

 Ok - before this degrades into the typical "ANSI is better" argument;
let's just capitulate and say "yes - all new code should be ANSI."

 Now; let's examine another possible situation, which we may want to
allow, or may want to not worry about.

 In my instance, I'm trying to write some rather portable software.
This software has to compile, on other platforms, with a non-ANSI
compiler... so, for that reason, I can only conditionally use ANSI-isms.

 In the past, I was using gcc -traditional on FreeBSD as a screening
test before the software was compiled on the non-ANSI host.   This would
quickly catch issues and give the developer writing the code an opportunity
to discover the issue before there had to be some massive "we've got
to fix this to compile on platform X" issue later.

 If we decide that gcc -traditional isn't going to work with our
system headers, then I can no longer employ that screening test (without
editing the headers, which is what I've tended to do for several
versions of FreeBSD now...)

 So - you see, I agree - all new code, when possible, should use the
appropriate ANSI  features... but I don't see a compeling reason to
want to disallow those situations where we can't.

 [By the way, the host in question is an RS/6000 with IBM's XLC compiler,
it has *way* too many bugs for us to use..., the non-ANSI version seems
a little more robust... and, we can't upgrade to newer ones because we're
stuck at AIX 3.2.5; for support reasons.]

	- Dave Rivers -


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message