From mboxrd@z Thu Jan  1 00:00:00 1970
Path: main.gmane.org!not-for-mail
From: Friedrich Dominicus <frido@q-software-solutions.com>
Newsgroups: gmane.emacs.help
Subject: Re: Newbie regexp question
Date: 31 Oct 2002 15:41:22 +0100
Organization: Q Software Solutions GmbH
Sender: help-gnu-emacs-admin@gnu.org
Message-ID: <87pttqwu0d.fsf@fbigm.here>
References: <mailman.1036012442.21874.help-gnu-emacs@gnu.org> <3DC13711.6F82B396@enea.se>
NNTP-Posting-Host: main.gmane.org
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Trace: main.gmane.org 1036075589 27981 80.91.224.249 (31 Oct 2002 14:46:29 GMT)
X-Complaints-To: usenet@main.gmane.org
NNTP-Posting-Date: Thu, 31 Oct 2002 14:46:29 +0000 (UTC)
Return-path: <help-gnu-emacs-admin@gnu.org>
Original-Received: from monty-python.gnu.org ([199.232.76.173])
	by main.gmane.org with esmtp (Exim 3.35 #1 (Debian))
	id 187GaY-0007H2-00
	for <gnu-help-gnu-emacs@m.gmane.org>; Thu, 31 Oct 2002 15:46:26 +0100
Original-Received: from localhost ([127.0.0.1] helo=monty-python.gnu.org)
	by monty-python.gnu.org with esmtp (Exim 4.10)
	id 187Gab-0007tz-00; Thu, 31 Oct 2002 09:46:29 -0500
Original-Path: shelby.stanford.edu!newsfeed.stanford.edu!skynet.be!skynet.be!newsfeed.online.be!newsfeed01.sul.t-online.de!newsmm00.sul.t-online.com!t-online.de!news.t-online.com!not-for-mail
Original-Newsgroups: gnu.emacs.help
Original-Lines: 146
Original-X-Trace: news.t-online.com 1036075242 00 7089 eHW9bxuSSXANHt 021031 14:40:42
Original-X-Complaints-To: abuse@t-online.com
X-Sender: 320004655587-0001@t-dialin.net
User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Common Lisp)
Original-Xref: shelby.stanford.edu gnu.emacs.help:106619
Original-To: help-gnu-emacs@gnu.org
Errors-To: help-gnu-emacs-admin@gnu.org
X-BeenThere: help-gnu-emacs@gnu.org
X-Mailman-Version: 2.0.11
Precedence: bulk
List-Help: <mailto:help-gnu-emacs-request@gnu.org?subject=help>
List-Post: <mailto:help-gnu-emacs@gnu.org>
List-Subscribe: <http://mail.gnu.org/mailman/listinfo/help-gnu-emacs>,
	<mailto:help-gnu-emacs-request@gnu.org?subject=subscribe>
List-Id: Users list for the GNU Emacs text editor <help-gnu-emacs.gnu.org>
List-Unsubscribe: <http://mail.gnu.org/mailman/listinfo/help-gnu-emacs>,
	<mailto:help-gnu-emacs-request@gnu.org?subject=unsubscribe>
List-Archive: <http://mail.gnu.org/pipermail/help-gnu-emacs/>
Xref: main.gmane.org gmane.emacs.help:3170
X-Report-Spam: http://spam.gmane.org/gmane.emacs.help:3170

Paul Cohen <paco@enea.se> writes:

> Hi all,
> 
> Thanks to everyone who has been kind to take time to answer my question! Jay's answer is definitely closest to solving my problem.
> 
> "Bingham, Jay" wrote:
> 
> > After seeing Mike and Friedrich's exchange on your question I decided to create my own solution, a general purpose function that allows the user to specify the start and end as regular expressions and supply a replacement for the text between them, and optionally by giving a numeric prefix argument to the function force it to replace the tags as well.
> 
> Neat.
> 
> > In the process I noticed that Mike's function has a logic flaw.  It will never find the case of a missing end tag and instead deletes that final start tag.
> 
> Ok.
> 
> > Here is my function if you want it.
> 
> Yes I do! :-)
> 
> >
> > (defun replace-between-regexp (start-re end-re repl-str &optional incl)
> >   "Replace the text between two regular expressions supplied as arguments.
> > With a numeric argument the regular expressions are included.
> > When called non interactively incl should be nil for non-inclusion and
> > non-nil for inclusion."
> >   (interactive "sStart regexp: \nsEnd regexp: \nsReplace between %s and %s with: \nP")
> >   (while (re-search-forward start-re nil t)
> >     (let ((beg (if incl (match-beginning 0) (match-end 0)))
> >           (end
> >            (progn
> >              (if (re-search-forward end-re nil t)
> >                  (if incl (match-end 0) (match-beginning 0))
> >                nil))))
> >       (if (not end)
> >           (error "Unmatched \"%s\" sequence at position %d" start-re beg)
> >         (delete-region beg end)
> >         (insert repl-str)))))
> 
> I have few comments/questions.
> 
> I tried the above function in my *scratch* buffer by writing it and then adding the following lines (with line numbers!):
> 
> 19. (replace-between-regexp "<!--Test-->" "<!--End of Test-->" "" 1)
> 20.
> 21. Pub
> 22. <!--Test-->
> 23. Foo
> 24. <!--End of Test-->
> 25. Bar
> 
> I then evaluated the function with C-j. This resulted in:
> 
> 19. (replace-between-regexp "<!--Test-->" "<!--End of Test-->" "" 1)
> 20.
> 21.
> 22. Pub
> 23. nil
> 24.
> 25. Bar
> 
> With the cursor on line 24. My comments/questions are:
> 
> 1) I understand that the "nil" on line 23 comes from the value of
> the last item in the function list, in this case "(insert
> repl-str)". But there is also a newline character is inserted after
> "nil". But I don't want either the "nil" or the extra newline
> character! 
I'm a bit lazy to answer all or suggest other things but here's are my
thoughts on that
Call it with (replace-betwe.... "<!-- End of Test" nil t)

than check before the output
(when repl-string (insert repl-string))


> 
> 2) Line 21 containing "pub" is moved forward to line 22. I guess
> this is just because I did C-j at the end of line 19 or?
Check the documentation of C-j
C-h k C-j gives it to you
> 
> 3) It would be nice if the cursor would return to its original
> position after running the command. I tried adding the
> "save-excursion" command after the "interactive" line in the function
>but it didn't work. The cursor still ended up on line 24.
You have to enclose it all in save-excursion

It looks like this than:
(defun replace-between-regexp (start-re end-re repl-str &optional incl)
  "Replace the text between two regular expressions supplied as arguments.
 With a numeric argument the regular expressions are included.
 When called non interactively incl should be nil for non-inclusion and
 non-nil for inclusion."
  (interactive "sStart regexp: \nsEnd regexp: \nsReplace between %s
and %s with: \nP")
  (save-excursion
    (while (re-search-forward start-re nil t)
      (let ((beg (if incl (match-beginning 0) (match-end 0)))
            (end
             (progn
               (if (re-search-forward end-re nil t)
                   (if incl (match-end 0) (match-beginning 0))
                 nil))))
        (if (not end)
            (error "Unmatched \"%s\" sequence at position %d" start-re beg)
          (delete-region beg end)
          (when repl-str (insert repl-str)))))))

Do not use C-j than but C-x C-e 

This is the documentation for C-j
Documentation:
Evaluate sexp before point; print value into current buffer.
So you will get nil after this run

C-x C-e runs


Documentation:
Evaluate sexp before point; print value in minibuffer.
With argument, print output into current buffer.

So the output will be put into the minibuffer.

> 
> 4) The idea to solve my problem with a lisp function is neat and I
> guess there are many situations where one would like to add ones own
> special purpose functions to Emacs. My question is: where is the
> suitable place to put them? In my .emacs file or in a separate
> file. What are the conventions? 
.emacs is change quite frequently. I don't think you self-written
functions will after development, therfor put them into an own file,
byte-compiler that file and load it from the .emacs file.

I've put all my stuff under .xemacs here and libraries I've installed
over time have found there way under ~/lib/elisp. It makes it quite
easy to update other systems such that XEmacs behaves like I want it
too.

It's IMHO nearly unavaiable that you customize your Emacs over time
and you feel like a fish on land while you are used to some things and
they are not there. 

Regards
Friedrich