Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site mit-eddie.UUCP
Path: utzoo!linus!decvax!genrad!mit-eddie!rms@prep.AI.MIT.EDU
From: rms@prep.AI.MIT.EDU
Newsgroups: net.emacs
Subject: [Donald.Stone@a.sei.cmu.edu: Gnu Bugfix]
Message-ID: <167@mit-eddie.UUCP>
Date: Thu, 13-Feb-86 12:07:58 EST
Article-I.D.: mit-eddi.167
Posted: Thu Feb 13 12:07:58 1986
Date-Received: Fri, 14-Feb-86 07:23:11 EST
Sender: daemon@mit-eddi.UUCP
Organization: MIT, Cambridge, MA
Lines: 27

From: rms@prep.AI.MIT.EDU (Richard M. Stallman)
Date: Wednesday, 12 February 1986 16:52:20 EST
From: Donald.Stone@a.sei.cmu.edu
To: rms%mit-oz@xx.lcs.mit.edu
Subject: Gnu Bugfix

Hi,

   Regarding the bug in Gnu on the Apollo
    (running out of Pure space)

   Pure is being allocated at A10000

   So in purecopy (and some other places, too) the test
[purecopy and mark_object in alloc.c, CHECK_IMPURE in lisp.h]

  if (XINT (obj) < (int) ((char *) pure + PURESIZE)
      && XINT (obj) >= (int) pure)

   always fails because XINT is sign extending the address.

   When I changed the XINTs to XUINTs, things worked.

   You'd know better than I would whether that fix will break 
   anything else.

        Don Stone (ds@a.sei.cmu.edu)