[emms-help] Why are maps declared with defconst?

[Pierre Neidhardt]

I wonder why maps are declared with defconst.

It seems to create issues with Evil, such as defining bindings with a
prefix.
I'm not completely sure why as I haven't dug into Evil internals.

Can we change the definition of the maps to using `defvar`?

[Yoni Rabkin]

Post by Pierre Neidhardt
I wonder why maps are declared with defconst.
It seems to create issues with Evil, such as defining bindings with a
prefix.
I'm not completely sure why as I haven't dug into Evil internals.
Can we change the definition of the maps to using `defvar`?

Yes, this would be a good change. How about the style shown in the Elisp
reference manual "11.6 Tips for Defining Variables Robustly"?

[Pierre Neidhardt]

Do you mean the following?

(defvar my-mode-map nil
DOCSTRING)
(unless my-mode-map
(let ((map (make-sparse-keymap)))
(define-key map "\C-c\C-a" 'my-command)
...
(setq my-mode-map map)))

[Yoni Rabkin]

Post by Pierre Neidhardt
Do you mean the following?
(defvar my-mode-map nil
DOCSTRING)
(unless my-mode-map
(let ((map (make-sparse-keymap)))
(define-key map "\C-c\C-a" 'my-command)
...
(setq my-mode-map map)))

yup

Personally, I can take it or leave it since it is obviously a kludge to
move the docstring closer to where it makes sense.

So moving to defvar with this or with the old style both sound fine to
me.

[Pierre Neidhardt]

Post by Yoni Rabkin
Personally, I can take it or leave it since it is obviously a kludge to
move the docstring closer to where it makes sense.

Agreed :)

[Pierre Neidhardt]

On a similar note, why do some maps inherit from text-mode-map?

(set-keymap-parent map text-mode-map)

I tried disabling it, it does not seem to change anything.

Thoughts?

[Yoni Rabkin]

Post by Pierre Neidhardt
On a similar note, why do some maps inherit from text-mode-map?
(set-keymap-parent map text-mode-map)
I tried disabling it, it does not seem to change anything.
Thoughts?

In cases like emms-playlist-mode I've always thought that it would be
cool to be able to freely annotate the buffer (I still do). Perhaps
that's what I was thinking... but who really knows since some of this
code is easily over 10 years old.

I wouldn't change it unless there is a compelling reason to do so.