From: Paul Richards <paul@originat.demon.co.uk>
Message-Id: <199607011810.TAA00562@originat.demon.co.uk>
Subject: Re: I *don't* want Paul's style guide.
To: new-httpd@hyperreal.com
Date: Mon, 1 Jul 1996 19:10:00 +0100 (BST)
In-Reply-To: <199607011755.NAA15643@hershey.ai.mit.edu> from "Robert S. Thau"
 at Jul 1, 96 01:55:54 pm
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Sender: owner-new-httpd@apache.org
Precedence: bulk
Reply-To: new-httpd@hyperreal.com

In reply to Robert S. Thau who said
> 
> It isn't just "other Apache-compatible servers" --- it's future
> versions of our own.  In particular, if any of the hard-core HTTP-NG
> proposals, with binary-encoded requests and responses, rather than
> RFC822-style headers, start to go places, we may want to change a lot
> of the core innards to accomodate that --- particularly in
> http_protocol.c, but also perhaps elsewhere.  That'll be easier if we
> haven't exported any protocol-dependent functionality to the modules
> in the meantime.
> 
> There may well be cases where something isn't supported and should be
> --- if it's just a case of some silly string-twiddling routine which
> people could just as easily write for themselves, the objections above
> obviously don't apply.  But in general, the question of what to export
> does require at least a little thought.

This is a totally separate issue to the style guide. Just because you
haven't put something in a header file doesn't stop people using it in
their own modules. We need to decide what the exportable interface of the
server is and document it so that people don't use the "innards" illegally
and prevent us changing them. Anything not exportable or required to
be global should be static in any case.

-- 
  Paul Richards, Originative Solutions Ltd.
  Internet: paul@netcraft.co.uk, http://www.netcraft.co.uk
  Phone: 0370 462071 (Mobile), +44 1225 447500 (work)