All (most) `_user` functions are handled here instead. To allow keyboard specific configuration, I've created `_keymap` functions that can be called by the keymap.c files instead.
My [Ergodox EZ Keymap](https://github.com/qmk/qmk_firmware/blob/master/layouts/community/ergodox/drashna/keymap.c#L297) is a good example of this, as it uses the LEDs as modifier indicators.
This borrows from @jola5's "Not quite neo" code. This allows me to maintain blocks of keymaps in the userspace, so that I can modify the userspace, and this is reflected in all of the keyboards that use it, at once.
The caveat here is that the keymap needs a processor/wrapper, as it doesn't like the substitutions. However, this is as simple as just pushing it through a define. For instance:
A bunch of macros are present and are only included on boards that are not the Ergodox EZ or Orthodox, as they are not needed for those boards.
Included is a custom macro for compiling my keyboards. This includes the bootloader target (`:teensy`, `:avrdude`, or `:dfu`), and keeps RGBLIGHT, AUDIO and/or FAUXCLICKY enabled, if it previously was (regardless of the rules file).
This uses the `layer_state_set_*` command to change the layer color, to indicate which layer it is on. This includes the default keymap, as well.
Since this is done via userspace, it is the same between all systems.
Additionally, there is a custom keycode to toggle layer indication. And all RGB keycodes disable layer indication by default, as well. This way, I can leave special effects doing when I want.
Also. I use `rgblight_sethsv` since it works with animation modes (that support it).
When Tap Dances are enabled, this layer has the ability to "spam" keypresses.
For instance, tapping the TD "1" twice causes the layer to hit "1" ever 1 second (appoximately). This is useful for auto-hotkeying skills (such as bone armor or devour).
Tappind once disables this, and switching layers temporarily disables this, until you switch back to the layer.
For critics that think this is cheating, search "diablo 3 num lock auto cast". This is just a simpler method, since I no longer own a normal (non QMK) numpad.
First, I have several files that are hidden/excluded from Git/GitHub. These contain everything needed for the macros. To hide these files, open `.git/info/exclude` and add `secrets.c` and `secrets.h` to that file, below the comments.
And this requires `KC_SECRET_1` through `KC_SECRET_5` to be defined in your `<name>.h` file to define the keycodes for the new macros.
And then, in the `process_record_user` function, assuming you have `return process_record_keymap(keycode, record)` here, you'll want to replace the "final" return with the following. Otherwise, you want to replace the `return true;` with `return process_record_secrets(keycode, record);`
Here, you want your `/users/<name>/rules.mk` file to "detect" the existence of the `secrets.c` file, and only add it if the file exists. To do so, add this block:
Additionally, if you want to make sure that you can disable the function without messing with the file, you need to add this to your `/users/<name>/rules.mk`, so that it catches the flag:
The reason that there was some issues before, is that JTAG was still enabled, and using some of the pins that the keyboard uses. Disabling JTAG (either by fuse, or modifying the matrix code for splits fixes the issue).
And for reference, if you want to go back to caterina, the default fuse settings I believe were: