Systemd is stable and can get the job done, I don’t love how it does it, but I accept that it’s a reasonable choice.
Something need/s/ed to replace X11. The fact that I still can’t autohotkey in Wayland is fucking insane. 9/10 of my current issues in Linux are Wayland-related. We’re being corralled into wayland as distros drop X11 and I don’t have time to keep chasing that fox.
Wayland enforces security to stop one process sniffing or injecting input into another. X11 basically has zero security boundaries. So certain things are easier in X11, but then again the price to pay is too much.
So things like virtual mouse positioning, autohotkey, screen capture are harder to do but not insurmountable. It requires protocols that the client / desktop understand and which are fit for purpose. Big desktops like GNOME & KDE Plasma do support xdg-desktop-portal which allows hotkeys to be registered with the desktop, and could potentially be what you need.
Stuff like ydotool, wtype exist and presumably there are others for this. Search on automation testing on Wayland and see.
Since it is a niche I’m not surprised if it wasn’t a critical path for Wayland adoption but clearly there are solutions that work for certain compositors and I assume people working on a more robust solution. But allowing one piece of software to sniff or inject input, or capture its content is an obvious security flaw and Wayland devs are right to put it front and centre. If a desktop is to allow those things then it should be through a security model instead of in the absence of one.
Systemd and Wayland have become ragebait for some people.
Systemd is stable and can get the job done, I don’t love how it does it, but I accept that it’s a reasonable choice.
Something need/s/ed to replace X11. The fact that I still can’t autohotkey in Wayland is fucking insane. 9/10 of my current issues in Linux are Wayland-related. We’re being corralled into wayland as distros drop X11 and I don’t have time to keep chasing that fox.
Wayland enforces security to stop one process sniffing or injecting input into another. X11 basically has zero security boundaries. So certain things are easier in X11, but then again the price to pay is too much.
So things like virtual mouse positioning, autohotkey, screen capture are harder to do but not insurmountable. It requires protocols that the client / desktop understand and which are fit for purpose. Big desktops like GNOME & KDE Plasma do support xdg-desktop-portal which allows hotkeys to be registered with the desktop, and could potentially be what you need.
I need mouse and keyboard control to do testing automation.
I want keyboard interception so i can make dozens of !faxnumber, !address, !bashloop, !bashwhile macros to use text replacement on
You can do this in every other OS. It wasn’t even a primary exploit, just something some devs decided to die on a hill over.
Wayland has been reasonably used for a decade, at what point would you consider unreasonbly difficult to transition into insurmountable
Stuff like ydotool, wtype exist and presumably there are others for this. Search on automation testing on Wayland and see.
Since it is a niche I’m not surprised if it wasn’t a critical path for Wayland adoption but clearly there are solutions that work for certain compositors and I assume people working on a more robust solution. But allowing one piece of software to sniff or inject input, or capture its content is an obvious security flaw and Wayland devs are right to put it front and centre. If a desktop is to allow those things then it should be through a security model instead of in the absence of one.
I know. My simple criteria is pretty simple: Does my computer boot and show me pretty pictures of kittens after it boots. If it does, I do not care.