From xemacs-m  Tue Mar 18 19:46:12 1997
Received: from crystal.WonderWorks.COM (crystal.WonderWorks.com [192.203.206.1])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id TAA17997
	for <xemacs-beta@xemacs.org>; Tue, 18 Mar 1997 19:46:10 -0600 (CST)
Received: by crystal.WonderWorks.COM 
	id QQchmd19318; Tue, 18 Mar 1997 20:46:11 -0500 (EST)
Date: Tue, 18 Mar 1997 20:46:11 -0500 (EST)
Message-Id: <QQchmd19318.199703190146@crystal.WonderWorks.COM>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
From: Kyle Jones <kyle_jones@wonderworks.com>
To: XEmacs Beta Mailing List <xemacs-beta@xemacs.org>
Subject: relocating allocator?
In-Reply-To: <199703190029.QAA14028@roy>
References: <199703190029.QAA14028@roy>
X-Mailer: VM 6.20 under 19.15 XEmacs Lucid (beta99)
X-Face: /cA45WHG7jWq>(O3&Z57Y<"WsX5ddc,4c#w0F*zrV#=M
        0@~@,s;b,aMtR5Sqs"+nU.z^CSFQ9t`z2>W,S,]:[+2^
        Nbf6v4g>!&,7R4Ot4Wg{&tm=WX7P["9%a)_da48-^tGy
        ,qz]Z,Zz\{E.,]'EO+F)@$KtF&V

Damon Lipparelli writes:
 > Does anyone remember the consensus on whether or not rel-alloc is a good
 > idea?  I seem to recall it being suspected for some crashes and performance 
 > problems, but I can't find a specific message.

    From: Jamie Zawinski <jwz@lucid.com>
    To: Kyle Jones <kyle@crystal.wonderworks.com>
    Subject: your rel_alloc lossage
    Date: Fri, 1 Apr 94 21:42:47 PST

    A side-effect of REL_ALLOC is that any call to malloc ***even in
    a library routine*** can cause buffer->text.beg to be relocated
    (and cause the ptr in text.beg, and only that ptr, to point to
    the new place.)

    So things in emacs that grab a copy of the old ptr in text.beg
    and hold on to it around calls to things which may malloc are an
    accident waiting to happen.

    I haven't investigated further since having this revelation.

	    -- Jamie

I never found out if someone did a sweep of the code to look for such
caching of b->text.beg.  Because I don't know, I have built XEmacs
with --rel-alloc=no since then.

