Message-Id: <Mutt.19961203150727.blong@hoohoo>
Date: Tue, 3 Dec 1996 15:07:27 -0600
From: blong@uiuc.edu (Brandon Long)
To: new-httpd@hyperreal.com
Subject: Re: your mail
References: <199612031616.QAA03110>
 <9612031618.aa13371@gonzo.ben.algroup.co.uk>
Mime-Version: 1.0
Content-Md5: qXOciphU820ax4+sbJBPyA==
Organization: Fiction L Networks
Sender: new-httpd-owner@apache.org
Precedence: bulk
Reply-To: new-httpd@hyperreal.com

On 12/3, Ben Laurie uttered the following other thing:
> So, what do I do about this kind of thing? I can put in a test easily enough,
> but someone still has to deal with conf.h...
> 
> Perhaps we should go on to configure with a special "unknown" OS, invite
> people to handle it in conf.h, and submit the patches, which we then rework
> into a proper set?
> 
> Or perhaps we should just modify Configure to encourage them to provide
> appropriate patches under their own steam?

Or, add a minimalist autconf type configure support to check for 
the easiest set of directives.  Ie, grep for sigjmp_buf in the
appropriate header files, look for specific header files to include
or not, the functions that can be defined or not to give hints, if not
fully working support.

Or, you might assume a basic SYSV if you don't know it already.
What percentage of operating systems will survive on sysv?  Though,
admins with those types of os's probably know enough to port the code.

BTW, why do you still use getwd?  Actually, looks like you only use
it mod_include.c, yet you have to define against it to getcwd for 7
different systems.

Brandon
-- 
 Brandon Long             "A man who could make so vile a pun would not
 MD6 Crash Test Dummy        scruple to pick a pocket."  
 Intel Corp, Oregon                 -- John Dennis
          I'm too low on the totem pole to speak for Intel.
		  http://www.uiuc.edu/ph/www/blong