From xemacs-m  Thu Sep  4 03:32:17 1997
Received: from inf.enst.fr (dz9pURgXwnf3BxiH1SS1UjPSyqVW+YBn@inf.enst.fr [137.194.2.81])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id DAA02180
	for <xemacs-beta@xemacs.org>; Thu, 4 Sep 1997 03:32:15 -0500 (CDT)
Received: from metheny.enst.fr (metheny.enst.fr [137.194.160.27])
          by inf.enst.fr (8.8.4/8.8.4) with ESMTP
	  id KAA19784 for <xemacs-beta@xemacs.org>; Thu, 4 Sep 1997 10:32:11 +0200 (MET DST)
Received: (from verna@localhost)
          by metheny.enst.fr (8.8.4/8.8.4)
	  id KAA05095; Thu, 4 Sep 1997 10:32:07 +0200 (MET DST)
To: xemacs-beta@xemacs.org
Subject: first match-data ever ...
X-Attribution: dv
X-Url: http://www-inf.enst.fr/~verna/
From: Didier Verna <verna@inf.enst.fr>
X-Face: 6o|eiKqaHN.ANh8HXDzntcWUOCg\]RsOd.ctvm~*y}Y^R&*a+Co,\s#=HWsw3x$b_n2kJ#g
 (7u?J^@^xP)f,jUF|0Z'J:|G/bMA5O12*b,7`-Q`=pKsCRIpso07.Y>YB2H{7`?u&yh;C_ZtLHfj<!
 $J=.i&Al'?,ax]MZd4tcm)_wF3$n*:f/lgS.;?Jr3T;Fl^q<qP**'tw
X-Face: |j}\)O|k##MrRz#VK$Jy=0r=3Qc,,a/Tr6*JQbE73dy17]<u3$*$]4O\1|h\|O\EDT9d$n+
 MTB{U&>2YcmW$9Z&H21e}#~#pgc>dn(is5Bv1l!{1re+Q9suKIOUmOqZs2>QMxHlR;;}kaGYA@HR3D
 C6
Mime-Version: 1.0 (generated by tm-edit 7.108)
Content-Type: text/plain; charset=US-ASCII
Date: 04 Sep 1997 10:32:06 +0200
Message-ID: <qyjsovllbsp.fsf@metheny.enst.fr>
Lines: 29
X-Mailer: Gnus v5.4.65/XEmacs 20.3(beta19) - "Kyiv"


	The problem introduced by my patch to emacs-version is that
this function seems to process a matching operation for the first time 
in the emacs session. Therefore, calling `save-match-data' will call in
turn `match-data' which returns an error because there 's no previous
match.[1]

	A possible workaround would be to check for the error in the
macro `save-match-dat' and preserve the previous mathcing context only 
if there is one, but we can't use `ignore-errors' since cl-macs is
loaded *after* replace :-(

	So finaly, I think that `match-data' shouldn't produce and
error when no matching operation was performed before. Couldn't it
return just nil instead ? I mean that people may want to preserve the
previous matching stuff, "just in case", and so it should be an error
to save a non existant context.

	Comments ? Anyway, I'm not going to touch the C code. I've done 
enough jokes for today ;-)

Footnotes: 
[1]  Steven, it's strange that you didn't get it ?

--
    /     /   _   _       Didier Verna        http://www-inf.enst.fr/~verna/
 - / / - / / /_/ /      E.N.S.T. INF C214        mailto:verna@inf.enst.fr
/_/ / /_/ / /__ /        46 rue Barrault        Tel.   (33) 01 45 81 80 72
                      75634 Paris  cedex 13

