From owner-freebsd-config  Tue Jan 28 13:07:04 1997
Return-Path: <owner-freebsd-config>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id NAA10136
          for config-outgoing; Tue, 28 Jan 1997 13:07:04 -0800 (PST)
Received: from perki0.connect.com.au (perki0.connect.com.au [192.189.54.85])
          by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id NAA10116;
          Tue, 28 Jan 1997 13:06:55 -0800 (PST)
Received: (from Unemeton@localhost) by perki0.connect.com.au id IAA14933
  (8.7.6h/IDA-1.6); Wed, 29 Jan 1997 08:06:52 +1100 (EST)
X-Authentication-Warning: perki0.connect.com.au: Unemeton set sender to giles@nemeton.com.au using -f
>Received: from localhost.nemeton.com.au (localhost.nemeton.com.au [127.0.0.1])
	by nemeton.com.au (8.8.5/8.8.5) with SMTP id IAA25280;
	Wed, 29 Jan 1997 08:00:09 +1100 (EST)
Message-Id: <199701282100.IAA25280@nemeton.com.au>
To: Michael Smith <msmith@atrad.adelaide.edu.au>
cc: config@freebsd.org, chat@freebsd.org
Subject: Re: Kernel config metasyntax 
In-reply-to: <199701281432.BAA12099@genesis.atrad.adelaide.edu.au> 
Date: Wed, 29 Jan 1997 08:00:09 +1100
From: Giles Lean <giles@nemeton.com.au>
Content-Type: text
Sender: owner-freebsd-config@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk


On Wed, 29 Jan 1997 01:02:23 +1030 (CST)  Michael Smith wrote:

> Er, you don't mean semantic checking, you mean content checking,
> correct?

Content checking is minimum (catch those typos) but semantic checking
for inconsistent options is useful as well.

> Insofar as is possible, sure.  I don't think, however, that the sort
> of checking that you describe is feasible, desirable or even
> necessary, given that the input isn't going to be user-supplied.

I don't mind you arguing 'desirable' or 'necessary' but feasible is
precisely what I am concerned about.  I dislike tools that don't
validate their inputs, and to build one that cannot validate seems
unwise.

(In fact checking is feasible with a tool such as you describe, but
would have to be a separate stage to parsing.)

The current design of 'config' where options are passed through as C
#defines is less than wonderful.  I took a support call once at HP for
someone who had set MAXDSIZ to be greater than 2^32.  Interesting
things happened but config should have caught it. :-(

Giles