Possible bug in replace-regexp?

Possible bug in replace-regexp?

[Moira]

Hi.

I've just attempted some regular expression search/replace in emacs -  
a thing I like to do in editors - and noticed some strange behaviour.  
I thought I'd rather ask here before posting a bug-report to early.

If my search contains a grouped quantor like this "myString\(.*?\)" a  
C-q,C-j newline in the replacement is added *before* the groups  
backreference, even if it's placed behind it.
Like in this replace: "myString\1[emacs C-q,C-j comes here]".

This should yield:
 >>>
...
some-stuff
myString more-stuff-in-the-myString-line
the-rest
...
<<<

Instead it causes this:
 >>>
...
some-stuff
myString
more-stuff-in-the-myString-line the-rest
...
<<<

I should add that this is Aquaemacs, a native GUI Emacs port for OS  
X. I haven't tried it the CLI version yet and I'm not 100% sure C-q,C- 
j is the way of putting a newline into the replacement.

Is this a bug (read: can you guys reproduce it?) and/or are there  
other ways to approach this problem?

Thanks for any advice and input on the subject.

Phillip

Re: Possible bug in replace-regexp?

[Peter Dyballa]

Am 21.11.2007 um 00:59 schrieb Moira:

> I should add that this is Aquaemacs

For me Aquamacs Emacs, the Carbon Emacs as distributed by David  
Reitter, does not work since last year. I have at least one other GNU  
Emacs too many. In none of them. not even the Cocoa Emacs.app, the  
same happens.

Your from expression is a bit unconventional: .*? – what can the ? do  
that .* are not able to do? Prevent greediness?


I'd recommend to repeat the test in a new instance of Aquamacs Emacs  
and then view with C-h l what exactly you did enter. If you can see  
from the *Help* buffer that no C-j or such was entered by you and the  
result of replace-regexp is still the same as described in your  
eMail, then it's time to choose 'Report Bug' from Aquamacs Emacs'  
Help menu.


BTW, Aquamacs Emacs has its own list:

	---------------------------- Info -----------------------------
	List Post: <mailto:macosx-emacs@email.esm.psu.edu>
	List Archives: <http://dir.gmane.org/gmane.emacs.macintosh.osx>

--
Mit friedvollen Grüßen

   Pete

Math illiteracy affects 7 out of every 5 Americans.

Re: Possible bug in replace-regexp?

[David Reitter]

On Nov 20, 11:59 pm, Moira <mo...@demetria.de> wrote:

> If my search contains a grouped quantor like this "myString\(.*?\)" a
> C-q,C-j newline in the replacement is added *before* the groups
> backreference, even if it's placed behind it.
> Like in this replace: "myString\1[emacs C-q,C-j comes here]".

I just tried to reproduce that with the latest version from CVS (based
on Emacs 22 CVS) and I can't reproduce this.

If there is still a bug, it's likely to be right in GNU Emacs (since
we don't mess around with Search/Replace functions)

Re: Possible bug in replace-regexp?

[Johan Bockgård]

Moira <moira@demetria.de> writes:

> If my search contains a grouped quantor like this "myString\(.*?\)" a
> C-q,C-j newline in the replacement is added *before* the groups
> backreference, even if it's placed behind it.
> Like in this replace: "myString\1[emacs C-q,C-j comes here]".

*? is non-greedy. \1 is the empty string.

-- 
Johan Bockgård