I’m putting all efforts to switch to Wayland on hold for another 2 years, minimum.
As we all know, writing a Wayland compositor from scratch is next to impossible. Luckily, there’s the wlroots project which aims to build a base library for this task. Basically every compositor except for GNOME and KDE uses it. (This is good! The less fragmentation, the better.)
wlroots is still very volatile, lots of changes with every release. Downstream users (i.e., the projects that write the actual compositor) have to constantly “chase” changes in wlroots. dwl, my favorite compositor at the moment, has recently switched their main
branch to target the wlroots git version instead of the latest release. My understanding is that they have to do this in order to keep up with wlroots (maybe I’m wrong).
Everything is volatile and a moving target.
Why does any of this matter for me? Because I have to eventually fork dwl or at least keep a patch set, and I don’t have the stamina to constantly fiddle with this stuff. I’m running my own X11 window manager, it’s highly specialized, and using just “some Wayland compositor out there” is a huge step backward that I’m not willing to take. I tried, it’s just painful and annoying with zero benefits.
So … it was fun experimenting with Wayland a bit, but I’m now back to waiting for things to settle down considerably.
#anhstrq