forked from forks/qmk_firmware
fixed the layer limit of LT() (#9492)
This commit is contained in:
parent
a8bb5840ad
commit
9ae8a45197
|
@ -19,7 +19,7 @@ These functions allow you to activate layers in various ways. Note that layers a
|
|||
|
||||
### Caveats :id=caveats
|
||||
|
||||
Currently, `LT()` and `MT()` are limited to the [Basic Keycode set](keycodes_basic.md), meaning you can't use keycodes like `LCTL()`, `KC_TILD`, or anything greater than `0xFF`. Specifically, dual function keys like `LT` and `MT` use a 16 bit keycode. 4 bits are used for the function identifier, the next 12 are divided into the parameters. Layer Tap uses 4 bits for the layer (and is why it's limited to layers 0-16, actually), while Mod Tap does the same, 4 bits for the identifier, 4 bits for which mods are used, and all of them use 8 bits for the keycode. Because of this, the keycode used is limited to `0xFF` (0-255), which are the basic keycodes only.
|
||||
Currently, `LT()` and `MT()` are limited to the [Basic Keycode set](keycodes_basic.md), meaning you can't use keycodes like `LCTL()`, `KC_TILD`, or anything greater than `0xFF`. Specifically, dual function keys like `LT` and `MT` use a 16 bit keycode. 4 bits are used for the function identifier, the next 12 are divided into the parameters. Layer Tap uses 4 bits for the layer (and is why it's limited to layers 0-15, actually), while Mod Tap does the same, 4 bits for the identifier, 4 bits for which mods are used, and all of them use 8 bits for the keycode. Because of this, the keycode used is limited to `0xFF` (0-255), which are the basic keycodes only.
|
||||
|
||||
Expanding this would be complicated, at best. Moving to a 32-bit keycode would solve a lot of this, but would double the amount of space that the keymap matrix uses. And it could potentially cause issues, too. If you need to apply modifiers to your tapped keycode, [Tap Dance](feature_tap_dance.md#example-5-using-tap-dance-for-advanced-mod-tap-and-layer-tap-keys) can be used to accomplish this.
|
||||
|
||||
|
|
Loading…
Reference in a new issue