Re: Emacs Lisp Programming Questions

[Kevin Rodgers]

David Combs wrote:
> In article <mailman.8228.1254921577.2239.help-gnu-emacs@gnu.org>,
> Kevin Rodgers  <kevin.d.rodgers@gmail.com> wrote:
> > clint.laskowski wrote:
> > > Hello, gnu.emacs.help. I have a few questions about programming in
> > > Emacs Lisp. I hope you can help. Here they are:
> > >
> > > 1. Is this a good place to ask questions about programming in Emacs
> > > Lisp, especially with regards to text processing? If there's a better
> > > place, I'd appreciate knowing.
> > >
> > > 2. I want to write an interactive Elisp program to remove sequential
> > > duplicate lines from a buffer. This buffer is not sorted, and it
> > > should not be sorted. The program should simply look for two
> > > sequential lines that are identical, delete one, and then move on to
> > > the next line and do it over until it reaches the end of the buffer.
> > >
> > > BUT, I do not want the answer to this problem (i.e., I don't want an
> > > Elisp answer) ... I want hints on how to program it. I want to learn
> > > the answer myself, if possible.
> > >
> > > Any ideas or pointers?
> > Start by writing a keyboard macro that does what you want.  Then you
> > can translate the commands invoked interactively into function calls.
>
> I guess "edit-kbd-macro" and grabbing what it types out
> would be the way to start that translation?

Yes, and use `C-h f' to find the precise calling convention for each function.

> > Note that you can search for sequential duplicate lines using a
> > regular expression.  
>
> Please show an example -- I myself will learn something
> from it.
>
> It'd be cool to be able to get a regexp to extend past the
> end of the line, yes?
>
> BUT WAIT!  THIS IS *EMACS* -- doesn't work by lines --
> whole buffer is just one huge long string, with newlines
> interspersed here and there.  So, how to refer to a newline
> as a plain old ordinary character, not "end of line" via "$"?

LFD aka control-J

> Yes, PLEASE, I'd REALLY like to see an example of such a regexp!

^\(.*\) LFD \1 LFD

(the spaces are just for readability)

> And if indeed that *can* be done, can you use that within
> a query-replace somehow?

Not query-replace (M-%), but query-replace-regexp (C-M-%):

C-M-%
^\(.*\) C-q C-j \1 C-q C-j RET
\1 C-q C-j RET

> (Or will that miss every other one or something, because
> if it sees two identical lines and replaces the 2nd
> by "", then what if there is a third one also identical --
> can't get the qr to back up to make a match?)

Yes, that could be a problem.  This might work:

C-M-%
^\(.*\) C-q C-j \(\1 C-q C-j \)+ RET
\1 C-q C-j RET

> Anyway, an example would sure be nice, however you do it,
> just so long as you do it via a regexp.  THANKS!

Hope that helps!

--
Kevin Rodgers
Denver, Colorado, USA