From xemacs-m  Thu Jul 31 04:54:15 1997
Received: from frege.math.ethz.ch (root@frege-d-math-north-g-west.math.ethz.ch [129.132.145.3])
	by xemacs.org (8.8.5/8.8.5) with SMTP id EAA14355
	for <xemacs-beta@xemacs.org>; Thu, 31 Jul 1997 04:54:14 -0500 (CDT)
Received: from midget.math.ethz.ch (vroonhof@midget [129.132.145.4]) by frege.math.ethz.ch (8.6.12/Main-STAT-mailer) with ESMTP id LAA10195 for <xemacs-beta@xemacs.org>; Thu, 31 Jul 1997 11:54:14 +0200
Received: (vroonhof@localhost) by midget.math.ethz.ch (8.6.12/D-MATH-client) id LAA27671; Thu, 31 Jul 1997 11:50:36 +0200
To: xemacs-beta@xemacs.org
Subject: Re: pkgdir ignored?
References: <E0wso94-00019Q-00@neptune.corp.comsat.com> <199707281210.IAA27980@news.smart.net> <m2pvs3cidb.fsf@altair.xemacs.org> <bywwmayyaa.fsf@midget.math.ethz.ch> <byvi1uyun9.fsf@midget.math.ethz.ch> <m2racf1zl7.fsf@altair.xemacs.org>
Mime-Version: 1.0 (generated by tm-edit 7.106)
Content-Type: text/plain; charset=US-ASCII
From: Jan Vroonhof <vroonhof@math.ethz.ch>
Date: 31 Jul 1997 11:50:35 +0200
In-Reply-To: SL Baur's message of 30 Jul 1997 23:54:12 -0700
Message-ID: <bylo2nh7o4.fsf@midget.math.ethz.ch>
Lines: 46
X-Mailer: Gnus v5.4.55/XEmacs 19.15

SL Baur <steve@xemacs.org> writes:

> Will they?  I see a lot of postings with Gnus 5.3 ...

For these people a package system is a disadvantage. If they only
upgrade the packages when they upgrade Emacs what will happen if they
are unbundled? This is why there must be a
Xemacs-20.x-the-world.tar.gz and extremely good instuctions for the
installer to please consider upgrading packages to.
 
> My perception of the major problem of upgrading is one of dropping in
> a new XEmacs without breaking existing site code.  Please correct me
> if I'm wrong.

I don't know abot the 'the'-part it, but is is a major problem yes.
Typical example. I prepared a fully tuned 19.15 setup. Our sysadmin
copied it to /usr/local. However the fact XEmacs now automaticly
includes subsirs of site-lisp means that it is picking up the old
auctex and gnus from the 19.14 install instead of the bundled ones.

> temacs will be included in the next round of the binary kits (it
> should have been included last time, but we didn't get around to it).

Then the super-adjusted puresize might be a problem?

> > Secondly it might gives you problems with PURESIZE (does this NEED to
> > be a compiletime constant?.
> 
> Yes.  It delimits a static C array.

May be it is time to rethink this whole stuff anyway. My next
killer-feature (after threading but before lexical scoping) would .elc
file that can mmapped.

For the mean time (on mmap capable systems): How about
 pure-size becomes a run time parameter to temacs.
 temacs starts by mmapping a pure-size size file and uses that as pure
 space
 on dump time it unmaps the file and dumps.
 xemacs on startup mmaps the file read-only (the file could be a
 special section in the binary?) at exactly the same location it was
 before.

Would this work?


