Xinput2 Backend

[Fuzzyzilla]

A backend implementation using X11's xinput version 2 extension.

Closes #10.

Unfortunately, this implementation relies heavily on heuristics to detect the class and capabilities of hardware. There is not - as far as I can tell - a better way to go about it, and no resource exists listing the set of standard values. While I have a sizeable collection of hardware to test on, the heuristics will very likely need updating as more hardware is tried (pads with several rings, pads with sliders, tools with roll/distance/wheel/slider, airbrushes..., etc.).

Once implementation is finished (not yet, some breaking changes are still pending), I would greatly appreciate testing help to ensure this detects the full capabilities on all hardware. I will probably put out a call for this when the implementation is ready prior to making a release :3

• Tool, Pad, Tablet hardware listing
• Tool events
• Pad events
• Hardware Add/Remove Events
• Timeout based Out events
• Mouse emulation
• Remove the use of Xinput 1.x APIs which "clients are requested to avoid mixing" with Xinput 2 (whoops)
  • Some necessary XI1 calls have no XI2 counterpart. I'm confused.
• Refactor (a lot of this was written while I was pulling my hair out in utter bafflement)

[Fuzzyzilla]

Good news, after digging through several driver implementations (xf86-input-libinput, xf86-input-wacom, xf86-input-evdev, xwayland), it seems things are more standardized than is exposed through public documentation, and the things that aren't standardized are presented through driver-specific properties. These have been implemented and it's all around more scrumptious, however the heuristics must unfortunately remain for when the drivers aren't available. Getting there!

[bokubeam]

Hi! I'm interested in using octotablet for a tileset creation tool I'm working on here: https://github.com/verkaboot/tessellate

I'm using X11, so I'm definitely interested in these updates. I would be happy to help if there's anything I can do.

[Fuzzyzilla]

Thank you for the interest ^^. I am very pleased to see this crate put to use :3
I got a bit caught in the weeds of implementation details on this PR, so I haven't touched this in a bit. I will say having someone in need of this feature will be great motivation for getting it done!

I believe I have most of the protocol and hardware/driver-specific quirks mapped out in my notes, which was the hard part.