feat(linux): implement hi res scroll

[rszyma]

Implemented REL_WHEEL_HI_RES and REL_HWHEEL_HI_RES for mwheel-* actions based on how my mouse sends hi-res scroll events.

From my testing, (mwheel-down 50 120) feels the same as before, while (mwheel-down 5 12) is much smoother (if an app supports hi res scroll events).

Also I don't know how, but it fixes #395

[jtroo]

Thanks for the PR!

That is quite strange that it fixes the bug, since it doesn't seem to change how passthrough of unhandled device events works - in this case the scroll events of a high res mouse.

Maybe it's a desktop thing where if the first type of scroll event from a given device is high-res, only high-res will be handled and low-res will be ignored. Or if low-res is sent first, then high-res will be ignored.

[jtroo]

It would be good to include the removal of the following text from config.adoc and kanata.kbd as part of this PR.

In Linux, not all desktop environments support the REL_WHEEL_HI_RES event, so kanata just doesn’t use it. Instead, a scroll happens when 120 or more distance units are accumulated. This may result in poor scrolling experience so in Linux it is recommended to use a distance value that is a multiple of 120.

  ;; ... In Linux, not all desktop environments
  ;; support the REL_WHEEL_HI_RES event so kanata just doesn't use it; instead a
  ;; scroll happens if 120 or more distance units are accumulated. This may result
  ;; in poor scrolling experience so in Linux it is recommended to use a distance
  ;; value that is a multiple of 120.

[rszyma]

Fixed wrong math operators.

[rszyma]

Now that I think about it, I didn't test if this didn't break normal mouse scroll wheel (without hi res scroll events). I don't have such mouse, so I need someone else to test it.

[rszyma]

Tested with 2 mice without high res scrolls: SteelSeries Rival 3 and Dell KM632. No issues.