From xemacs-m  Thu Sep  4 18:20:44 1997
Received: from mercury.Sun.COM (mercury.Sun.COM [192.9.25.1])
	by xemacs.org (8.8.5/8.8.5) with SMTP id SAA09304
	for <xemacs-beta@xemacs.org>; Thu, 4 Sep 1997 18:20:44 -0500 (CDT)
Received: from Eng.Sun.COM ([129.146.1.25]) by mercury.Sun.COM (SMI-8.6/mail.byaddr) with SMTP id QAA29582; Thu, 4 Sep 1997 16:20:13 -0700
Received: from kindra.eng.sun.com by Eng.Sun.COM (SMI-8.6/SMI-5.3)
	id QAA17345; Thu, 4 Sep 1997 16:20:10 -0700
Received: from xemacs.eng.sun.com by kindra.eng.sun.com (SMI-8.6/SMI-SVR4)
	id QAA05763; Thu, 4 Sep 1997 16:20:07 -0700
Received: by xemacs.eng.sun.com (SMI-8.6/SMI-SVR4)
	id QAA21486; Thu, 4 Sep 1997 16:20:07 -0700
Date: Thu, 4 Sep 1997 16:20:07 -0700
Message-Id: <199709042320.QAA21486@xemacs.eng.sun.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
From: Martin Buchholz <mrb@Eng.Sun.COM>
To: "Stephen J. Turnbull" <turnbull@turnbull.sk.tsukuba.ac.jp>
Cc: xemacs-beta@xemacs.org
Subject: Eventual success Debian 1.3 Linux 2.0.29
In-Reply-To: <m0x6ZZx-00000wC@turnbull.sk.tsukuba.ac.jp>
References: <m0x6ZZx-00000wC@turnbull.sk.tsukuba.ac.jp>
X-Mailer: VM 6.33 under 20.3 "Vienna" XEmacs  Lucid (beta14)
Reply-To: Martin Buchholz <mrb@Eng.Sun.COM>

>>>>> "SJT" == Stephen J Turnbull <turnbull@turnbull.sk.tsukuba.ac.jp> writes:

SJT> Build report for Debian GNU/Linux 1.3 kernel 2.0.29
SJT> Sorry, haven't had time to do more than play Tetris once.  Applause.
SJT> Took some doing to get a clean build.

SJT> First, the configure script seems to be hosed.  I've noticed in a
SJT> number of reports that it does not record the location of X Windows
SJT> header files.  I've only checked to make sure on a few random reports
SJT> for Linux systems, but the configure script lossage is _definitely_
SJT> _not_ just on my system.  It does do the right thing for the X11 libs.

I respectfully suggest that any system that has SOME, BUT NOT ALL of
the X11 header files in /usr/include/X11 is totally hosed.  

What were these guys thinking?

"Oh, we'll provide some header files in /usr/include so that your X11
app looks like it might actually build, but instead fails mysteriously."

``Almost'' only counts in horseshoes and hand grenades.

Unfortunately, both Linux and HP-UX 10.20 are known to have this
problem, so we'll likely have to find a workaround.  Unortunately++,
this detection is in the autoconf 2.12 code, not in configure.in
itself.  Someone with access to one of the broken systems will have to 
figure out how to do this.  

In the meantime, just do 
configure --x-includes=/usr/X11R6/include
on Linux and HP-UX and ????

Hhhhhhhhhhhmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm.....................

Oh, all right....  I surrended.  Hack, hack, hack...

The next 20.3 beta will contain an (as usual) completely untested
patch to deliberately avoid /usr/include and /usr/lib for X stuff if
the X11R6 alternatives are available, even if they appear at first
sight to be correct.

Beta testers: Please try the next beta on those platforms without
--x-includes or --x-libraries flags.

Martin

