Wayland-Protocols 1.5 Released With Tablet Support

[phoronix]

Phoronix: Wayland-Protocols 1.5 Released With Tablet Support

The latest release of the Wayland Protocols package is now available...

http://www.phoronix.com/scan.php?page=news_item&px=Wayland-Protocols-1.5

[rabcor]

I want to know more about tablet support..

Does it support pressure sensitivity?
How configurable is it (like xsetwacom?)
Does it support off-brand tablets like huion and yiynova through digimend or comparable means? (Using digimend on evdev for a yiynova, really wish I could have libinput instead!)

Give me all the infos! what's in? What's planned? what's in the future?

[garnacho]

I want to know more about tablet support..

Does it support pressure sensitivity?

I'm unsure if you mean about pressure changes being detected by applications, or whether the pressure applied is configurable. If it's the former, yes sure! along with tilt/distance/rotation/... as long as the tablet/tool provide the information. If it's the latter, that's up to the compositor to implement.

How configurable is it (like xsetwacom?)

Pretty much like with other devices' configuration, configurability is left up to compositors. That said, there is plenty of room for configurability, it probably won't be available through a CLI tool though.

Does it support off-brand tablets like huion and yiynova through digimend or comparable means? (Using digimend on evdev for a yiynova, really wish I could have libinput instead!)

AFAIK those should already work out of the box for the most part.

Give me all the infos! what's in? What's planned? what's in the future?

Actually making this widespread

[rabcor]

Sounds like it's leaving way too much up to the compositor... also about testing off-brand tablets on libinput... there should be configurability through CLI or a settings file somewhere (like xinput/xset or like xorg.conf, but for libinput specifically, preferably agnostic of environment/compositor)

I just so happen to have one, I'm using evdev right now with X.

How would I test it with libinput on X?

(Libinput at least does not automagically detect off-brand tablets after the digimend drivers were installed, and neither does evdev, it had to be specifically added like so: https://digimend.github.io/support/howto/drivers/evdev/ )

Configurability of libinput being tied to compositor leaves too much up to chance and too many openings for inconsistencies, this does not sound good at all. It's too much to expect compositor developers to develop configuration methods for every possible type of device libinput supports and testing it with different device vendors (wacom tablets and off-brand tablets don't behave the same way, I think only waltop behaves in a way comparable to wacom, everyone else does a different thing)

[garnacho]

Sounds like it's leaving way too much up to the compositor... also about testing off-brand tablets on libinput... there should be configurability through CLI or a settings file somewhere (like xinput/xset or like xorg.conf, but for libinput specifically, preferably agnostic of environment/compositor)

Well... it sounds like it's leaving too much to the compositor because it's compositors which decide how to implement it (if) at all. Wayland protocols don't dictate implementation specifics, they just define the messaging so one client and the compositor can communicate. In this case the vast majority of clients are interested in receiving input events, which is what the protocol caters for. The amount of "clients" that want to modify the global configuration are far more limited, most likely to 1, and I suspect you'll be disappointed at what DEs/compositors individually choose.

I just so happen to have one, I'm using evdev right now with X.

How would I test it with libinput on X?

(Libinput at least does not automagically detect off-brand tablets after the digimend drivers were installed, and neither does evdev, it had to be specifically added like so: https://digimend.github.io/support/howto/drivers/evdev/ )

xf86-input-libinput 0.18 is already out with stylus support, if you need tweaking xorg.conf.d you can try filing a bug to xf86-input-libinput.

Configurability of libinput being tied to compositor leaves too much up to chance and too many openings for inconsistencies, this does not sound good at all. It's too much to expect compositor developers to develop configuration methods for every possible type of device libinput supports and testing it with different device vendors (wacom tablets and off-brand tablets don't behave the same way, I think only waltop behaves in a way comparable to wacom, everyone else does a different thing)

The differences between brands, if any, are meant to be evened out by libinput. Granted, whoever develops this in a compositor will probably need *some* drawing tablet around.