FWIW, given peoples complained about hysteresis algo latency, I'm also
poking around with alternative hysteresis algos. One possible idea that
came to my mind is to use the detection code posted here to figure the
maximum wobbliness length (along each axis separately), then simply
ignore moveme
On Sun, Feb 18, 2018 at 01:09:22PM +0300, Konstantin Kharlamov wrote:
> For the purposes it seems to work fine — it's marked RFC because I see a
> small oddness, and I think it's better to ask someone more
> acknowledgable in libinput codebase.
thanks, much appreciated. I was hoping I get to this