From xemacs-m  Fri Feb  7 08:11:02 1997
Received: from alphatech.com (erebus.alphatech.com [198.112.236.2])
	by xemacs.org (8.8.5/8.8.5) with SMTP id IAA07689
	for <xemacs-beta@xemacs.org>; Fri, 7 Feb 1997 08:11:00 -0600 (CST)
Received: from venus.alphatech.com by alphatech.com (4.1/SMI-4.1)
	id AA11374; Fri, 7 Feb 97 09:06:39 EST
Received: from pochacco.alphatech.com by venus.alphatech.com (4.1/SMI-4.1)
	id AA20554; Fri, 7 Feb 97 09:14:12 EST
Received: by pochacco.alphatech.com (SMI-8.6/SMI-SVR4)
	id JAA08308; Fri, 7 Feb 1997 09:05:24 -0500
Sender: greg@pochacco.alphatech.com
From: greg@alphatech.com (Greg Klanderman)
To: xemacs-beta@xemacs.org
Subject: Re: M-p in isearch broken
References: <kig6805ueqs.fsf@jagor.srce.hr>
	<rxsu3noyl57.fsf@midnight.ecf.teradyne.com>
	<kigwwskucgv.fsf@jagor.srce.hr> <m24tforhcm.fsf@altair.xemacs.org>
Reply-To: greg@alphatech.com
Date: 07 Feb 1997 09:05:23 -0500
In-Reply-To: Steven L Baur's message of 07 Feb 1997 03:22:17 -0800
Message-Id: <ugd8uc4spo.fsf@pochacco.alphatech.com>
Lines: 65
X-Mailer: Gnus v5.2.40/XEmacs 19.15


OK, are we still agreed that its broken as to cycling the search ring
correctly?  I do think search-ring-update=nil is the behavior I'd
like, but I think this is even more broken than cycling incorrectly:

do xemacs -q and enter the following into *scratch* (between delimiters)

--- *scratch * ---

one

two

three

four

------------------

Now, do the following commands (messages in minibuffer shown on right,
followed by my comments)

M-<
C-s one <RET>
C-s two <RET>
C-s three <RET>
C-s four <RET>
C-s                  I-search:                // OK
M-p                  Search:                  // OK
M-p                  Search: two              // what???  should be "four" IMO
C-s                  Failing I-search: two    // OK
C-s                  Failing I-search: three  // what??? now its "three"??
C-s                  Wrapped I-search: three  // OK - wraps and finds "three"

If you use this feature often you see it jumping all over in the
search ring as you C-s within isearch, even when you think you've
selected the entry you want with M-p and M-n.

FWIW this seems less broken than it was in 19.13.  I'm running 19.15b4.

Greg



>>>>> "sb" == Steven L Baur <steve@miranova.com> writes:
sb> 
>> Yes, that's it.  But you have to press M-p twice to begin cycling
>> through the ring -- and that's a bug (as a side effect, the routine
>> *remembers* that you pressed it twice, and actually moves it twice --
>> *duh*).  We have RET to move us into `Search' mode.
sb> 
sb> C-h v search-ring-update
sb> 
sb> search-ring-update's value is nil
sb>   -- a variable declared in Lisp.
sb> 
sb> Documentation:
sb> *Non-nil if advancing or retreating in the search ring should cause search.
sb> Default nil means edit the string from the search ring first.
sb> 
sb> -- 
sb> steve@miranova.com baur
sb> Unsolicited commercial e-mail will be billed at $250/message.
sb> 
sb> 

