Migrating to Wayland

[okratitan]

This is going to be addressed sometime in the future but I wanted to go ahead and get a create a spot where discussion can happen as we move along. Do we want to have support for both X and Wayland backends? Or do we want to phase our X and phase in Wayland? Or do we want to remain with X? If X is going to remain one way or another is the logical progression forward to add compositing with X first and then add a Wayland backend? Do we want to just leave X without compositing and move to Wayland?

[mainiak]

Even if you have X as primary, it should be possible to switch to Wayland backend as future desktops should move away from X system.

I am using SwayWM on Wayland and it seems fast - I know, very subjective, plus I don't have X11 to compare it, nor I know how would I do any such comparison.

What I know, I wanna try Fyne Desktop, but I am not going to install X11 just for it. At least on my primary system.

[andydotxyz]

We do need to move to Wayland - but let's get Fyne apps working on that first ;) The decision about compositor probably depends on how long the wayland move takes, we have already seen limitations of non-composited X11 which will become more obvious over time.

Probably need to complete #88 first.

[andydotxyz]

Now the refactor is complete we can make a new internal/wayland package that could be used to boot up a wayland instance. The cmd/fynedesk/main_unix.go file would determine which to boot.

It's possible that the x11/win package may be referenced for XWayland support, but I think it may not be that simple ;).

[andydotxyz]

With Fyne 2.5 pretty much completing Wayland support in the toolkit this should be moved as a priority I think. Here os what I think the todo list is:

• [ ] Add wayland support for the main compositor
• [ ] Duplicate / port all window management code from X11 to Wayland
• [ ] Support XWayland to help migration

And later:

• [ ] Deprecate and remove X11 support