Touchpad taps don't work with tap-hold mods like real touchpad clicks (haptic or mechanic) on both Win and macOS

[newsve]

Since I started using my thumb with the touchpad, tapping it is way nicer than clicking. Before I preferred tapping too but now, it's so much faster and super precise. But tapping somehow doesn't seem to work when modifiers are held (mods from tap-hold, both tap-hold-press, tap-hold-release), .

[Didn't know if this should be an issue because mouse matters are so complex or this isn't supported anyway, so feel free to move]

(defcfg
  concurrent-tap-hold yes
  rapid-event-delay 15
  chords-v2-min-idle-experimental 5
  log-layer-changes no
  process-unmapped-keys yes
)

(defsrc a s g)

(deflayer one
  (tap-hold-release 0 200 a lsft)
  (tap-hold-release 0 200 s lctl) ;; the mac version has a lmet here
  (tap-hold-release 0 200 g lalt) 
)

Some example use cases which all work with real clicks (haptic or mechanical) flawlessly :

1. Try control-tap on a link in Chrome (so hold a and tap the touchpad)
   => this works but not with the first but second tap
   => Edit: or to be precise, you need to hold a an eternity before a tap would be registered, way more than the 200ms tap-hold timeout
2. In PowerShell/MS Terminal or another terminal emulator, try to mark something to get it into Windows clipboard which is holding shift, or if you want to select some square you would need to hold alt and shift together and then double tap and hold to drag
   a + double tap and hold (that typical touchpad drag) and moving the finger/pointer => nothing
   a + g + double tap and hold and moving the finger/pointer => nothing

Again, this works with real clicks.

The debug log just shows the mods as press/release loop but not the taps.

Edit and PS: Look that both, tap to click and the double tap and hold to drag are enabled on OS-level

[jtroo]

It working with real clicks is interesting, because unless you're intercepting the mouse events (which is possible in Linux/Interception, unsure macOS) Kanata has no way to know that a tap-hold-release or similar should trigger early hold activation on mouse action.

In that sense the problem behaviour is the default expected behaviour.