Bug: Odd behavior when splitting tap-hold to press and release

[oblitzitate]

Requirements

• I've searched platform-specific issues, issues and discussions to see if this has been reported before.

Describe the bug

I wanted to split up tap-hold to its press and release forms (i.e. combining virtualkeys with tap-hold) for better control so I can experiment with it (e.g. conditional press or release of a tap-hold).

The issue is it doesn't emit the tap or hold action in an expected way. Instead I have to double tap a key in order to emit the tap action, whereas the hold action is normally the one that emits (even on tap) and it will often emit on timeout (rather than on key release).

Relevant kanata config

(defalias
  g (multi
      (on-press press-vkey tp)
      (on-release release-vkey tp)
  )
)
(defvirtualkeys
  tp (tap-hold 0 990 g b)
)

To Reproduce

1. Apply the config
2. Experiment with tapping or holding the g key

Expected behavior

I was expecting it to behave like a normal tap-hold action.

Kanata version

e.g. kanata 1.6.1-prerelease-1

Debug logs

No response

Operating system

Linux 6.8.8-arch1-1

Additional context

No response

[gerhard-h]

the output is strange simulate.
With t:100 there not even is output.

[jtroo]

Ah yes, this is based on the way tap-hold is designed. When a tap hold is pending, no other actions are processed until something terminates its ending state. On a normal key, release would end the pending state. But virtual keys rely on a special action activation to release, which itself is blocked because of the pending state.