Re: Intelligent history cycling

Re: Intelligent history cycling

[Stefan Monnier <foo@acm.com>]

>>>>> "Kai" == Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:
> I use this:
> (mapcar
>  (function
>   (lambda (x)
>     (define-key x (kbd "ESC p") 'previous-complete-history-element)
>     (define-key x (kbd "ESC n") 'next-complete-history-element)))
>  (list minibuffer-local-completion-map minibuffer-local-isearch-map
>        minibuffer-local-map minibuffer-local-must-match-map
>        minibuffer-local-ns-map))
> I wish there was a default keybinding for this.

I also use it just like you.  Actually, if you look at simple.el you'll
see that there is code that does it as well, but it's commented out.
I wish we could turn it on again.

I think Richard didn't like it for some reason.  Maybe it was because
it fails when nothing matches or because C-x C-f M-p only lists the
previous file(s) in the same directory.
I've actually changed mine such that M-p only does search if the
user has already typed something (so that C-x C-f M-p behaves just
like it does now).  Also M-p reverts back to non-searching if the
search fails.


        Stefan

Re: Intelligent history cycling

[Kai Großjohann]

"Stefan Monnier <foo@acm.com>" <monnier+gnu.emacs.help/news/@rum.cs.yale.edu> writes:

> I also use it just like you.  Actually, if you look at simple.el you'll
> see that there is code that does it as well, but it's commented out.
> I wish we could turn it on again.

Maybe it's enough to find free keys?  Hm.  Lessee...  One possibility
is to use C-n/C-p for next/prev history element, and M-n/M-p for
next/prev complete history element.

C-c C-n/C-c C-p are still free.

C-M-n/C-M-p could be used, since these commands don't appear
necessary in the minibuffer.

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)

Re: Intelligent history cycling

[Stefan Monnier]

> > I also use it just like you.  Actually, if you look at simple.el you'll
> > see that there is code that does it as well, but it's commented out.
> > I wish we could turn it on again.
> Maybe it's enough to find free keys?

I don't see a need for it.  M-n and M-p work just fine.  You can always
get the old (i.e. current) behavior with C-a C-k M-p and C-a C-k M-n
(although in most cases you don't need to since M-p and M-n are mostly used
before doing anything else (before modifying the prompt) in which case
my code is careful to not do a history search but just the good
old "fetch next/prev history item").


	Stefan

PS: your "From:" header is giving problems because it's split on a second
    line just between Kai and Grossjohann.  Sometimes I see a space
    between the two, and sometimes not, depending on what piece of code
    it goes through.  Furthermore, it triggers a harmless bug in my mail
    reader :-(

Re: Intelligent history cycling

[Richard Stallman]

    I think Richard didn't like it for some reason.  Maybe it was because
    it fails when nothing matches or because C-x C-f M-p only lists the
    previous file(s) in the same directory.

I prefer the current behavior of M-p and M-n.
We have other commands to search the history list.