Date: Fri, 28 Jun 1996 12:27:29 +0100
Message-Id: <199606281127.MAA27827@cadair.elsevier.co.uk>
To: new-httpd@hyperreal.com
Subject: Re: KNF document
In-Reply-To: <199606271631.MAA06031@hershey.ai.mit.edu>
References: <199606271631.MAA06031@hershey.ai.mit.edu>
From: Paul Richards <p.richards@elsevier.co.uk>
Sender: owner-new-httpd@apache.org
Precedence: bulk
Reply-To: new-httpd@hyperreal.com

Robert S. Thau writes:
 >   I finally found the KNF document. It's *very* close to the general
 >   consensus
 > 
 > Sorry.  It makes several recommendations on which there is no
 > explicit consensus, and with which I frankly disagree.  These include
 > use of the _P() macro and K&R-style declarations for function arguments,
 > to allow use of obsolete pre-ANSI compilers, putting enum values in 
 > all caps, and the obscure rules about order of declaration of local
 > variables, which again seem based on obsolete technology (e.g.,
 > declaring the largest local variables first, to shrink the stack 
 > frame for comiplers too dumb to reorder the things on their own).

I wasn't suggesting adopting it wholesale, I meant to adopt it as the
discussion document and then revise it, I'd agree with most of your points
above. I'm having a hard time remebering what rule 3a1 is supposed to bloody
look like and if we just revise an actual document I think we'll reach
consensus quicker since the tallies seem to reflect the *general* KNF
approach.

I'll prepare another draft that's more up to date and in tune with
current opinion.