>>>> "Torsten" == Torsten Duwe
<duwe(a)caldera.de> writes:
Adrian> Above does not crash my (emacs-version) "XEmacs 21.4 (patch 6)
Adrian> \"Common Lisp (candidate 1)\" [Lucid] (i586-pc-win32) of Thu Nov
Adrian> 22 2001 on D5DC120J"
Discussion moved to xemacs-beta(a)xemacs.org!
Torsten> Hmm,
Torsten>
http://www.xemacs.org says:
Torsten> [...]
Torsten> Current Releases
Torsten> Stable branch:
Torsten> 21.1.14
Torsten> Gamma branch:
Torsten> 21.4.5
This will need to be updated by the release engineer (hint, hint).
Torsten> Beta branch:
Torsten> 21.5.3
Torsten> [...]
Torsten> and anonftp agrees with this :-(
Will need to be updated too.
Adrian> I get this instead:
Adrian> Signaling: (wrong-type-argument bufferp 1) ...
Torsten> Yes, it probably depends on lisp-{,dis}union and compiler
Torsten> optimizations. Maybe your (CVS ?) version also has
Torsten> additional checks. Did you use cygwin or some "other"
Torsten> compiler ?
system-configuration
"i586-pc-win32"
means Windows native.
I use MS VS 6 on
Microsoft Windows 2000 [Version 5.00.2195]
Torsten> Anyway, one of the many versions affected is XEmacs 21.4
Torsten> (patch 4) "Artificial Intelligence" [Lucid]
Torsten> (i386-Caldera-linux, Mule) of Thu Oct 25 2001 on
Torsten> proliant.caldera.de
Torsten> I strongly doubt M-x emacs-bug adds more relevant info,
How will you ever know without trying it? :-)
Torsten> but I can tell you this: "gcc version 2.95.2 19991024
Torsten> (release)"'s assembler output for decode_buffer at -O3 on
Torsten> an i386-linux-elf shows that it has re-ordered the lisp
Torsten> object deref in front of the allow_string-test for speed
Torsten> purposes.
Torsten> This crashes on said platform if decode_buffer is ever
Torsten> passed a non-pointer object. (A workaround is using -O1
Torsten> or less when compiling).
Torsten> Torsten
Torsten> P.S.: sorry for continuing a discussion on xemacs-patches
Torsten> ...
No need to appologise, just don't do it, OK?
xemacs-patches is for patches and followups to the patches.
I was walking a tight line myself, following up to xemacs-patches.
I wanted to indicate we should discuss this before commiting.
Best regards,
Adrian
--
Adrian Aichner
mailto:adrianï¼ xemacs.org
http://www.xemacs.org/