[07:40] <duflu> anpok: Does libinput have any special support for smooth two finger scrolling or is it just button 4/5 emulation?
[07:45] <anpok> duflu: yes it has two finger scrolling support
[07:46] <duflu> anpok: Does it have /good/ scrolling support (not emulating a wheel) ? :)
[07:47] <anpok> duflu: it offers two information.. the emulated number of ticks.. and a finer grained (and hopefully properly scaled) movement vector
[07:47] <duflu> anpok: So something like a normalized float vector relative to the touchpad dimensions?
[07:48] <anpok> remember.. when we introduced libinput I default to the second ccase
[07:48] <anpok> but that turned out to be problematic with downstreams..
[07:48] <anpok> so we need to expose that via a real axis in mir client
[07:48] <duflu> Heh. Compatibility or good software. Choose one
[07:48] <anpok> then we could use it gtk and qt..
[07:49] <duflu> anpok: My feeling is that the nice implementations of the world roughly map the touchpad height to scrolling the height of the screen. So if we could move in that direction with more precision I think it would be nice
[07:50] <anpok> oh..
[07:55] <duflu> You could abstract that using a float (or fixed point) motion vector
[08:01] <anpok> duflu: libinput does have device specific scaling of the measured scroll gestures
[08:01] <duflu> Hmm, methinks we need a nice demo client to play with it