About undefined keys in keypad mode

[norris-young]

If keypad mode is designed for executing commands, with a leader keymap(as shown below) without defining a, I think it should do nothing. However, a SPC a inserts a letter a in normal state instead. I really feel confused about this behaviour. Is there any variable to disable it?

(meow-leader-define-key
  ;; SPC j/k will run the original command in MOTION state.
  '("j" . "H-j")
  '("k" . "H-k")
  ;; Use SPC (0-9) for digit arguments.
  '("1" . meow-digit-argument)
  '("2" . meow-digit-argument)
  '("3" . meow-digit-argument)
  '("4" . meow-digit-argument)
  '("5" . meow-digit-argument)
  '("6" . meow-digit-argument)
  '("7" . meow-digit-argument)
  '("8" . meow-digit-argument)
  '("9" . meow-digit-argument)
  '("0" . meow-digit-argument)
  '("/" . meow-keypad-describe-key)
  '("?" . meow-cheatsheet))

[eshrh]

I think this is a pretty reasonable request. Please rebuild your meow on the latest commit, because in dcc705b I added the variable meow-keypad-self-insert-undefined, which you would want to set to nil.

I don't think that this is good for most users though, so I erred on the side of caution and left the default value to t, so this does not change default behavior.

[fbergroth]

I don't think that this is good for most users though, so I erred on the side of caution and left the default value to t, so this does not change default behavior.

I'm curious, in what scenarios do people want the self-insert fallback?

[eshrh]

Honestly now that you make me think about, I couldn't tell you. I guess I just accepted how it works and took it for granted. Do you think we should set the default to nil? Or do you think we should remove the option and make this the only behavior?

The reason this happens in the first place isn't a sparse/normal keymap issue like you might expect, but rather that the leader internally looks up keys under the standard C-c mode-specific-map (and also adds dispatch commands there as well). Maybe this was unintentional?

[fbergroth]

I can't think of any use case for the self-insert fallback either.

I like to keep options to a minimum if possible, so I think we should remove it unless we find a compelling reason to keep it.

[DogLooksGood]

Well, in motion state. If you bind x to a command, and SPC x will execute the original command on x? I mean, without any setup.

[norris-young]

Unless you bind x to H-x with meow-leader-define-key, it will still do self-insert.
And in my cases, motion state is usually used in read-only buffers, so I will choose the binding for original command on x instead of this self-insert.