Meow affecting binding to mode-specific-map in custom leader map?

[okamsn]

I have adjusted meow-keypad-leader-dispatch and meow-keymap-alist to point to a custom leader keymap. In this custom keymap, I bind "i" to mode-specific-map.

So far, when I try something like SPC i g to mean C-c g, Meow intercepts this as if I had pressed KEYPAD g for C-M-. Pressing SPC g correctly becomes C-M-. Pressing SPC i SPC g translates to C-c g.

Besides changing meow-keypad-leader-dispatch and meow-keymap-alist, what else needs to be done to stop meow-keypad from intercepting presses for mode-specific-map?

[DogLooksGood]

The default behavior of g is prefix next input with C-M-, unless you disable it, it's always translated as C-M-. It's specified with variable meow-keypad-ctrl-meta-prefix. BTW, having another custom leader map isn't a good idea with Meow, unless you are using a distribution like doom which provides you a leader map out-of-box.

[okamsn]

> The default behavior of g is prefix next input with C-M-, unless you disable it, it's always translated as C-M-.

I don't see that every g is always translated to C-M-. For example, if I press SPC a g, that works as expected and runs the bound command. I suspect that the problem with SPC i g is that SPC i is bound to mode-specific-map. So I ask, are meow-keypad-leader-dispatch and meow-keymap-alist the only variables that need to be changed?

The behavior is not what I expected after changing the variables, so I am asking about it.

> BTW, having another custom leader map isn't a good idea with Meow, unless you are using a distribution like doom which provides you a leader map out-of-box.

Yeah, I'm trying to combine my existing Evil leader map with the Meow keypad with minimal changes.

I see what you are saying now. Thank you for the information. It looks like if the keymap is detected to have a binding using M-, the keypad will use g as C-M-. If it doesn't detect that, then g works normally.

I'll update my leader map to adapt to that.

I saw that, on Emacs 29, the (lookup-key keymap "^[") in meow--keypad-has-sub-meta-keymap-p returns non-nil even when the found keys are bound to nil. If that is also the behavior on other Emacs versions, I would like to make a PR for treating that case as having no keys using meta.

[DogLooksGood]

Well, the perfect solution would be having a real simulation with some low-level translation. But I haven't figured out how to do it yet.