Re: kmacro.texi

[storm@cua.dk (Kim F. Storm)]

Richard Stallman <rms@gnu.org> writes:

> Reading your proposed changes to kmacro.texi made me aware that C-x
> C-k has become a sort of special command mode, in which many ordinary
> commands have special meanings.  That is against the spirit of Emacs,

I looked for 'spirit' in the index of the Lisp ref, and didn't find it.

> and going against the spirit of Emacs in something that is rather
> obscure and only rarely used is a bad thing.

Why do you think this is rarely used -- I use it _often_.

And C-x e e e ...  already does something very similar.
Nobody has complained about that so far.

And C-u C-SPC C-SPC C-SPC ...

(Ok, I'm biased, as I added those short-cuts too)

>
> So I want to get rid of this mode-like feature as it now stands.
>
> However, there is a way it could be implemented, giving similar
> results, that would be consistent with the spirit of Emacs: to make
> this a major mode.  One command could create a buffer which represents
> the macro ring, and switch to it.  That buffer could use a special
> major mode in which single-character commands do things like navigate
> through this ring.

Lets consider C-x C-k C-p C-p C-p C-k which rotates the macro ring
three places and executes the resulting head macro.

While the user enters C-x C-k C-p ... the echo area shows the current
macro at the head of the macro ring after each C-p, so in effect it is
much like a major mode -- just using the echo area as a single line
window into the macro ring.

Instead, you want to open a different window to show the macro ring
and select another macro in that window.

However, very often, the macros I write refer to data in several
windows on the screen, i.e. they contain things like C-x o.  E.g. to
navigate between a *grep* buffer and source windows, and back again.

With your proposal, I have to mess up the window layout just to rotate
the macro ring -- I would consider that to be a HUGE drawback compared
to the current functionality.

But there are some nice possibilities with your proposed interface,
so I'll think more about it.

>
> For instance, if it makes one line per keyboard macro, then moving
> point in this buffer could have the effect of rotating the ring.  That
> way, C-p and C-n in this buffer would have the desired effect on the
> macro ring, merely because they move point (as they usually do) in the
> buffer that represents the ring.
>
> Kim, could you work on these changes?

Yes.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk