tap_interrupted=True parameter does not work in context of LayerTap[BUG]

[regicidalplutophage]

Describe the bug

The combination of prefer_hold=True and tap_interrupted=True is an extremely useful setting that allows to both minimize typos and doesn't feel as sluggish to use as prefer_hold=False. This behavior is default in ZMK which is en exemplary firmware in terms of "ergonomic" features. Unfortunately, when applied in context of LayerTap, whether through KC.MT or KC.HT, pressing and releasing a key within tap_time produces weird results.

Sidenote: Since KC.MT is now just an alias for KC.HT, I think it should be deprecated completely.

To Reproduce Steps to reproduce the behavior:

• Setup and configuration of the affected parts of the firmware (avoid copy-pasting the entire configuration if possible)

TEST1 = KC.MT(KC.F, KC.MO(3), prefer_hold=True, tap_interrupted=True, tap_time=1000)
TEST2 = KC.HT(KC.F, KC.MO(3), prefer_hold=True, tap_interrupted=True, tap_time=1000)
TEST3 = KC.HT(KC.F, KC.LSFT, prefer_hold=True, tap_interrupted=True, tap_time=1000)

• Setup and configuration of peripherals
• Input: keys pressed, ...

1.

• TEST2 is pressed
• KC.F is pressed
• KC.F is released within tap_time
• TEST2 is released Result: stuck f

2.

• TEST2 is pressed
• TEST3 is pressed
• TEST3 is released within tap_time
• TEST2 is released Result: f, which is the tap output of TEST3 located on the same layer as TEST2

TEST2, TEST3 and KC.F are all located on the same layer. TEST1 is omitted for brevity as it produces the same result as TEST2

Expected behavior Expected results are:

1. enter located on layer 3
2. backslash located on layer 3

Debug output If applicable, add debug output from the serial console to help explain your problem.

Additional context Add any other context about the problem here.

[regicidalplutophage]

This is #737 but more condensed and in current version of KMK.