From xemacs-m  Thu Jan 23 12:08:38 1997
Received: from venus.Sun.COM (venus.Sun.COM [192.9.25.5])
          by xemacs.org (8.8.4/8.8.4) with SMTP
	  id MAA16264 for <xemacs-beta@xemacs.org>; Thu, 23 Jan 1997 12:08:37 -0600 (CST)
Received: from Eng.Sun.COM ([129.146.1.25]) by venus.Sun.COM (SMI-8.6/mail.byaddr) with SMTP id KAA01399; Thu, 23 Jan 1997 10:08:05 -0800
Received: from kindra.eng.sun.com by Eng.Sun.COM (SMI-8.6/SMI-5.3)
	id KAA00166; Thu, 23 Jan 1997 10:07:50 -0800
Received: from xemacs.eng.sun.com by kindra.eng.sun.com (SMI-8.6/SMI-SVR4)
	id KAA01945; Thu, 23 Jan 1997 10:07:34 -0800
Received: by xemacs.eng.sun.com (SMI-8.6/SMI-SVR4)
	id KAA09925; Thu, 23 Jan 1997 10:07:34 -0800
Date: Thu, 23 Jan 1997 10:07:34 -0800
Message-Id: <199701231807.KAA09925@xemacs.eng.sun.com>
From: Martin Buchholz <mrb@Eng.Sun.COM>
To: The Pope <lazarus@mind.net>
Cc: xemacs-beta@xemacs.org
Subject: 20.0-b91 Build Failure
In-Reply-To: <199701231515.HAA13804@binah.mind.net>
References: <199701231515.HAA13804@binah.mind.net>
Reply-To: Martin Buchholz <mrb@Eng.Sun.COM>
Mime-Version: 1.0 (generated by tm-edit 7.100)
Content-Type: text/plain; charset=US-ASCII

>>>>> "Pope" == The Pope <lazarus@mind.net> writes:

Pope> Well, this build failure is no big deal...just the pure space being
Pope> exhausted.  I'm gonna go try it again, but here's my relevant data:

Steven, it's probably not a good idea to drop the purespace right
before a release.  Pure usage is highly dependent on architecture and
built-in features, and the computation in puresize.h is only an
approximation.  If you're going to shave it close, you need to update
and improve the pure usage calculation.  Unfortunately, even if you
wanted to do such a thing, it would  be difficult, since you can't
experimentally derive things like values when LONGBITS=64 without a
64-bit OS.

My recommendation is to raise the puresize macros in puresize.h back
to their previous values.  The only reason puresize.h has the current
values is because someone ran out of purespace sometime in the past.

Currently puresize.h is a maintenance headache.  A user who wants to
have a site-load.el will need to learn about the --puresize flag.  If
we were sufficiently clever, we could figure out a way to get rid of
puresize.h, and determine the puresize needed dynamically.

Martin

