Re: regexp to collapse multiple blank lines into one

[Neon Absentius]

On Mon, Oct 10, 2005 at 10:35:01AM -0600, Kevin Rodgers wrote:
> B. T. Raven wrote:
> > "Neon Absentius" <absent@sdf.lonestar.org> wrote in message
> > news:mailman.10556.1128865819.20277.help-gnu-emacs@gnu.org...
> >>The following seems to work
> >>
> >>C-M-% \(^C-qC-j\)+ RET \1 RET
> >>
> >>Provided of course that your lines are really blank and do not contain
> >>any spaces, tabs etc.
> >
> > Thanks F.L., P.D., N.A. Neon's solution seems the easiest to me. I need
> > this only because I can't get control chars into the replacement string.
> > The replacement can be a bare RET but you can't replace with less than
> > nothing, i.e. replacing a regexp with nothing and then backspacing to
> > get rid of the blank line.
> 
> Why can't you use the same technique to put a control character into the
> replacement string as Neon suggests for the search string/regexp
> (namely, C-q)?
> 

Actually that also works.  The reason I used "\1" is that I had
already grouped the regexp so it saved me typing.

> I'll admit, I don't understand why that solution works: it appears to
> replace the matched sequence of LFD characters with itself, instead of
> replacing it with just a single LFD character (or nothing at all,
> depending on the desired result).
> 

I don't understand what you don't understand :) My solutions replaces
one or more occurences of a regexp with a single occurence.  Perhaps
you are worried abour efficiency? that if there is a single occurence
I replace it with itself?  In that case you are right.  A better
solution would be better:

C-M-% \(^C-qC-j\)\{2,\} RET \1 RET

-- 
 There is no national science just as there is no national
multiplication table; what is national is no longer science.
         -- Anton Checov