bug#74903: [PATCH] Add function to show table of keys when describing keymap variable

[Mekeor Melire <mekeor@posteo.de>]

[-- Attachment #1: Type: text/plain, Size: 2209 bytes --]

On December 16, 2024 6:28:58 PM GMT+01:00, Robert Pluim <rpluim@gmail.com> wrote:
>>>>>> On Mon, 16 Dec 2024 17:57:14 +0200, Eli Zaretskii <eliz@gnu.org> said:
>
>    >> From: Mekeor Melire <mekeor@posteo.de>
>    >> Date: Sun, 15 Dec 2024 23:14:15 +0000
>    >> 
>    >> It'd be nice if users could opt in to have C-h v some-map RET show a
>    >> nice table of key bindings for that keymap.  I wrote a function that can
>    >> be added to help-fns-describe-variable-functions so that this happens.
>    >> 
>    >> What do you think?
>    >> 
>    >> Alternatively, we can also add a key binding to help-mode so that users
>    >> can easily "switch" from describe-variable to describe-keymap, e.g. per
>    >> "K":
>    >> 
>    >> C-h v message-mode-map RET K
>
>    Eli> We have "C-h b" and "C-h m" which show the bindings, albeit not by
>    Eli> keymap.  Is that not enough for some reason?
>
>"C-h b" shows all the bindings split out by major and minor mode
>already. I guess we could add the name of the relevant keymap as well,
>but the people who need to know the name of the keymap generally know
>enough about Emacs to figure it out (eg via "C-h k").
>
>    Eli> Adding yet another help command related to key bindings needs to be
>    Eli> justified, since we already have at least two with overlapping
>    Eli> functionalities.
>
>    Eli> What do others think about this?
>
>I really donʼt see the need for more functionality here.
>
>Robert
>-- 

I get the impression that you talk about M-x describe-keymap RET which does already exist.

That is not what I proposed. What I proposed is two things:

1. Allow users to opt-in to have C-h v foo-keymap RET show a table of key bindings in addition to the formatting as S-expression. This fits well into the existing help-fns mechanism. The patch I submitted implements a prototype.

2. In help-mode, bind a single-char-key like "K" to a command that will allow users to easily switch from describing a variable as variable to describing it as keymap a la describe-keymap.

I proposed (2.) as an alternative to (1.) but my main proposal is (1.). Both approaches have a similar objection though.


[-- Attachment #2: Type: text/html, Size: 3490 bytes --]