From xemacs-m  Wed Sep 24 21:33:27 1997
Received: from kurims.kurims.kyoto-u.ac.jp (kurims.kurims.kyoto-u.ac.jp [130.54.16.1])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id VAA12191
	for <xemacs-beta@xemacs.org>; Wed, 24 Sep 1997 21:33:19 -0500 (CDT)
Received: from orion.kurims.kyoto-u.ac.jp (orion.kurims.kyoto-u.ac.jp [130.54.16.5]) by kurims.kurims.kyoto-u.ac.jp (8.8.5/3.4W2) with SMTP id LAA14061; Thu, 25 Sep 1997 11:32:31 +0900 (JST)
Received: (from petersen@localhost) by orion.kurims.kyoto-u.ac.jp (SMI-8.6/3.5Wbeta) id LAA10834; Thu, 25 Sep 1997 11:32:31 +0900
To: Martin Buchholz <mrb@Eng.Sun.COM>
Cc: XEmacs Beta List <xemacs-beta@xemacs.org>
Subject: Re: thing.elc (b23)
References: <lb4t7bz2ff.fsf@orion.kurims.kyoto-u.ac.jp> <199709240623.XAA09601@xemacs.eng.sun.com>
X-Face: fOOYdM>Ct-+jZ!MfKWRl?5e=(j4]xHE~<%D)$=FuN"@nP]"fi8stnK9>1fW>1HR[fj\=i%q0'l2G]0[H'R*m(fm^Og;iG>lBmr{anp!LG[)PD}g^XbG.(@oIi8;y)8+C"HV5}8NQ`HR7~P{+$AEEPo,N>,IwIbd:Dq6\"Bi#Aq\^lJ~dmwnUkc-Xb{k3\dZnF7j,$enH=Ybr<9v,Kzi-T|urefU:c{$pC)T5yvnRO0T+D&Z9{B8ulAd5X#c?Z|}vv^qhBjEj
X-Emacs: 20.3 "Sarajevo" XEmacs  Lucid (beta23) with mule
Mime-Version: 1.0 (generated by SEMI MIME-Edit 0.86 "Naka-Tsurugi")
Content-Type: multipart/mixed;
 boundary="Multipart_Thu_Sep_25_11:32:30_1997-1"
Content-Transfer-Encoding: 7bit
From: Jens-Ulrik Holger Petersen <petersen@kurims.kyoto-u.ac.jp>
Date: 25 Sep 1997 11:32:30 +0900
In-Reply-To: Martin Buchholz's message of "Tue, 23 Sep 1997 23:23:30 -0700"
Message-ID: <lb67rqcedd.fsf@orion.kurims.kyoto-u.ac.jp>
Lines: 64
X-Mailer: Gnus v5.5/XEmacs 20.3(beta23) - "Sarajevo"

--Multipart_Thu_Sep_25_11:32:30_1997-1
Content-Type: text/plain; charset=US-ASCII

>>>>> "Martin" == Martin Buchholz <mrb@Eng.Sun.COM> writes:

>>>>> "Jens-Ulrik" == Jens-Ulrik Holger Petersen <petersen@kurims.kyoto-u.ac.jp> writes:
    Jens-Ulrik> `thing-symbol' in "util/thing.elc" is giving me the
    Jens-Ulrik> following backtrace, when I do

    Jens-Ulrik> M-: (thing-symbol (point)) RET

(at the end of a buffer as Martin pointed out).

    Martin> This'll fix it.

    Martin> --- lisp/utils/thing.el.old
    Martin> +++ lisp/utils/thing.el

Thanks.


I still don't understand why the old "thing.el" doesn't give an error,
while its .elc file does.  This looks like a Elisp bug, or am I
missing something?

I mean consider the file


--Multipart_Thu_Sep_25_11:32:30_1997-1
Content-Type: application/octet-stream; type=emacs-lisp
Content-Disposition: attachment; filename="char-test.el"
Content-Transfer-Encoding: 7bit

(defun synt-test ()
  (char-syntax nil))

(synt-test)

--Multipart_Thu_Sep_25_11:32:30_1997-1
Content-Type: text/plain; charset=US-ASCII


save it and byte-compile it.

The source ".el" file loads without error, but the ".elc" file gives
an error:

---- Start of included text -----------------------8<--- cut here -------------
Load error for /home/petersen/emacs/test/char-test.elc:
(wrong-type-argument characterp nil)

Failed to load 1 of 1 file
(char.elc)
	Thu Sep 25 11:14:47 1997	Buffer `test<emacs>'
---- End of included text -------------------------8<--- and here -------------

Can people reproduce this?

-- 
Jens
---
Does the name Pavlov ring a bell?

--Multipart_Thu_Sep_25_11:32:30_1997-1--

