From xemacs-m  Mon May 19 12:32:59 1997
Received: from piglet.splode.com (h-205-217-243-176.netscape.com [205.217.243.176])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id MAA25830
	for <xemacs-beta@xemacs.org>; Mon, 19 May 1997 12:32:58 -0500 (CDT)
Received: (from friedman@localhost)
	by piglet.splode.com (8.8.5/8.8.5) id KAA06401;
	Mon, 19 May 1997 10:32:17 -0700
Date: Mon, 19 May 1997 10:32:17 -0700
Message-Id: <199705191732.KAA06401@piglet.splode.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
From: Noah Friedman <friedman@splode.com>
To: neal@ctd.comsat.com
Cc: xemacs-beta@xemacs.org
Subject: mmap hpux configure
Reply-To: friedman@splode.com
In-Reply-To: <neal@ctd.comsat.com> Mon, 19 May 1997 09:01:25 -0400
References: <E0wTS45-0000pf-00@data.ctd.comsat.com>

>As I expected, 20.3b1 fails to detect mmap on hpux9 (and I expect
>hpux10).  The test for mmap is overly strict.  I have had no luck
>convincing the autoconf gods to change this.  Maybe we can all try? 
>
>The mmap test should be changed to check for different levels of
>support, rather than pass/fail.

The truth is, the autoconf mmap test is obsolescent.  You can't merely
determine whether mmap "works" or not, because many systems have different
filesystem types where mmap works for some but not others.  You really have
to test at runtime to see if mmap works with the particular file you're
trying to map.

About all autoconf can really tell you reliably is that the mmap call
exists.

I don't think anyone is maintaining autoconf at the moment.  No one has had
any luck getting in contact with David MacKenzie lately.  I may take over
maintenance just to fix serious problems, but nothing is definite yet.

