~leon_plickat/lavalauncher

1

Lavalauncher, strange behavior with modifiers and layer=top

Details
Message ID
<CAJ1y1imWt5HMSbJVLUt8PQYicPOX=_tMz20EmNEZ99aYEn2L9w@mail.gmail.com>
DKIM signature
missing
Download raw message
Hello Leon,

First off, thank you for work on Lavalauncher.
I am trying to configure the application, but I'm seeing some very
weird behavior which may be a bug or it may be that I have
misunderstood something.

My system is a recently installed Arch Linux, with a few relevant
packages listed below:
[jens@yoga ~]$ pacman -Qs
'^(lavalauncher|cairo|wayland|wayland-protocols|sway)$'
local/cairo 1.17.4-5
    2D graphics library with support for multiple output devices
local/lavalauncher 2.1.0-2
    A simple launcher panel for Wayland
local/sway 1:1.6.1-1
    Tiling Wayland compositor and replacement for the i3 window manager
local/wayland 1.19.0-1
    A computer display server protocol
local/wayland-protocols 1.21-1
    Specifications of extended Wayland protocols

Ideally, I'd like to configure the launcher to be "a hidden bar which
sits above windows on hover", as described in the manual.

However, I end up in one of two "broken" states:
 (1)  if there is *no* button configured with a binding that includes
a keyboard modifier, the launcher will not react on any input (hover,
click, touch, ...)
 (2)  if there *is* a button configured with a binding that includes a
keyboard modifier AND layer=top or overlay, the launcher will react
properly to mouse and touch events, but all keyboard input will be
"swallowed" by Lavalauncher! The only exception being Sway key
bindings such as to change workspace, but in this state, I cannot get
keyboard input focus to any other running application. Thus I can't
even kill lavalauncher without going to another PTY and killing it
from there (I have to use SIGKILL / -9 at this stage to make it go
away).

(3) The only configuration which renders a working state, is when
layer is not set (=bottom), AND there is at least one button
configured with a binding that includes a keyboard modifier.

For a demonstration of the issue (including my configuration), see
screencast below showing these states in order (3), (1), (2).
https://photos.app.goo.gl/1bqaMi8aUmVKvxSK9

If it is of any importance, the laptop I'm using is a 2-in-1
convertible which has touchpad, touchscreen and can be folded into
tablet mode.

Please let me know if you need any more information.
Thank you very much for taking your time.

Best regards,
Jens
Details
Message ID
<CDWZUPX24MOT.3837UJIEFUAUO@m5>
In-Reply-To
<CAJ1y1imWt5HMSbJVLUt8PQYicPOX=_tMz20EmNEZ99aYEn2L9w@mail.gmail.com> (view parent)
DKIM signature
missing
Download raw message
On Mon Aug 30, 2021 at 6:31 PM CEST, Jens Hagerman wrote:
> However, I end up in one of two "broken" states:
> (1) if there is *no* button configured with a binding that includes
> a keyboard modifier, the launcher will not react on any input (hover,
> click, touch, ...)

Huh, that sounds like bug that I already fixed some time ago. I'll
investigate.

From your screencast I see that you are either on the latest release
or master branch. Could you perhaps try the "lavalauncher-next" branch?
It's where LavaLauncher development happens currently and I think it
has the fix for your problem (beware that the config format changed though).

> (2) if there *is* a button configured with a binding that includes a
> keyboard modifier AND layer=top or overlay, the launcher will react
> properly to mouse and touch events, but all keyboard input will be
> "swallowed" by Lavalauncher! The only exception being Sway key
> bindings such as to change workspace, but in this state, I cannot get
> keyboard input focus to any other running application. Thus I can't
> even kill lavalauncher without going to another PTY and killing it
> from there (I have to use SIGKILL / -9 at this stage to make it go
> away).

This is a known issue and something I can't really do anything about
sadly. There is no way for a surface in the layer shell to just get
the modifiers when necessary without grabbing the keyboard entirely.

TBH I am contemplating removing support for keyboard modifiers in the
future, as it never correctly worked and I don't think that will
change any time soon.




Friendly greetings,
Leon Henrik Plickat
Reply to thread Export thread (mbox)