Invert Y tilt axis when using UCLogicTiltReportParser

[jkcdarunday]

Introduction

While testing UCLogic tilt support using Huion RTP-700 with Krita and MyPaint on Linux, I've observed that the Y tilt axis is inverted.

Description

These changes inverts the Y tilt axis when a TiltTabletReport is generated by the UCLogicTiltReportParser.

Notes

• This assumes that the tilt axes for other devices have no issues when on Linux (making this specific to the UCLogic parser and unlikely somewhere else like libevent)
• Haven't tested on Windows/Mac (If it works properly on other platforms, then it's possible that the problem lies with the libevent integration instead)

[terencode]

Thanks, that's also what I noticed with my Kamvas pro 20. I will give it a try soon.

[terencode]

I modified the parser responsible for my tablet directly because it seemed to redefine itselfs the X, Y tilts :
https://github.com/terencode/OpenTabletDriver/blob/d44bdeb2a74482c970931f167532c73e5973483f/OpenTabletDriver.Configurations/Parsers/Huion/GianoReport.cs#L21
Tilt is correct now.
It would be interesting to test on other devices if this is parser specific or needs to be changed globally.

[AbstractQbit]

What are the values shown in the debugger? They should be negative for tilt to the left or up, positive for tilt to the right or down (for X and Y respectively). Two huion tablets that I had my hands on (hs610 and hs611) as well as wacom tablets report tilt correctly, so this issue is either device-specific or something is wrong with how we inject tilt on linux. In any case flipping Y axis shouldn't be the default

[terencode]

What are the values shown in the debugger? They should be negative for tilt to the left or up, positive for tilt to the right or down (for X and Y respectively). Two huion tablets that I had my hands on (hs610 and hs611) as well as wacom tablets report tilt correctly, so this issue is either device-specific or something is wrong with how we inject tilt on linux.

Without the Y flipped, I get positive values when tilting up, so device-specific ?

In any case flipping Y axis shouldn't be the default

Yeah, my modification was just a dirty fix for testing.

[gonX]

This seems sort of sensible but I wonder if we should be fixing this another place in OTD, or with an attribute in the device configuration file.

Also HS610 and HS611 does not use a tilt-compatibler parser (at the moment), though they should very likely be using the UCLogicTiltReportParser as used here.

I'd like for this to have more testing, especially on the HS610/HS611 as this'll determine which course of action is best. @AbstractQbit can you confirm HS610/HS611 tilt axises are correct on configs (not this PR) if you use UCLogicTiltReportParser?

[AbstractQbit]

Back at home now, seems like I was wrong. Just checked hs611 and it does have Y axis inverted, so it's possible that a lot of UClogic tablets report tilt that way. Can't check hs610 because I can't find it.

[AkiSakurai]

https://github.com/DIGImend/digimend-kernel-drivers/blob/0b445150ef491cb52174aef484f6873083770171/hid-uclogic-params.c#L472
According to DIGImend, it is flipped on device using v2 protocol.

[gonX]

Sorry for not getting back to this earlier. This needs to be rebased to the latest version of master before we can proceed with merging.

[sezanzeb]

Rebasing worked locally without conflicts.

For the huion kamvas 16 (2021) this needs to be adjusted in the GianoReport class.

I think this doesn't take into account that the input can be rotated. When I hardcoded the inversion into the GianoReport class, I had to invert the X axis, not the Y axis. My tablet is rotated by 180°.