Maarten van Gompel

Indeed, Cmp + . should give access to : _ is under Cmp + , .. I can't reproduce the problem, all seems fine here, are you on the latest...

That suggests that a call to `zwp_virtual_keyboard_v1_modifiers()` preceeds this one to `zwp_virtual_keyboard_v1_keymap()`: https://github.com/jjsullivan5196/wvkbd/blob/8106d7606d5d0ea24d23b3f0fadeafe7e6e528f6/keyboard.c#L695 , which would indeed be invalid. I haven't seen this occur in other situations. Could you try...

That's odd indeed, the logs and your debug statement suggest everything is done in the proper order so I think the error lies with the compositor. As far as I...

The behaviour might be explained if the compositor doesn't accept the keymap for some reason, like invalid size like in the case of #35?, but that supposed to be already...

I provided some writeup in https://github.com/Smithay/smithay/issues/1355#issuecomment-2028681912, I think the issue is in `xkb::Keymap::new_from_fd()` in [xkbcommon-rs](https://github.com/rust-x-bindings/xkbcommon-rs/) (which Smithay uses, and Niri in turn uses Smithay).

> What about F11 & F12? > Also Ins key Those indeed aren't mapped on any of our current layouts. You can always make your own layout though that includes...

I'm not sure what causes this, wvkbd reacts to both touch and pointer events (which are separate things in wayland). Try to see what your pen outputs using a tool...

I'm not sure what you mean, if by disable you mean hide wvkbd, then yes, send signal SIGUSR2 to hide and SIGUSR1 to show again (or SIGRTMIN to toggle).

> My understanding is that the wm/compositor should send these signals (eg. SIGUSR1/2, etc...) on entering a text field. No, any other process can send it (such as a gesture...

From a terminal you can send the signal like: hide: ``kill -USR1 $(pidof wvkbd-mobintl)`` show: ``kill -USR2 $(pidof wvkbd-mobintl)`` toggle: ``kill -34 $(pidof wvkbd-mobintl)`` You can't hide/show wvkbd from wvkbd...