So I've slowly been trying to resurrect hikari -- a cwm(1)-like compositor for #wayland, written by @acmelabs but hasn't received any love.
I know that darts is preferred. but... fuck that.
I've moved it over to git (history intact), and mirrored things here:
https://github.com/ThomasAdam/hikari
In trying to drag this thing free of dust, I've been working on a few things:
1) Use meson to build
2) Upgrade wlroots from the support 0.15 version to 0.18
Guess how point 2 is going...
Abysmally.
https://github.com/ThomasAdam/hikari/tree/ta/wlroots-upgrade
So far, I'm at 0.16, working on 0.17. The end goal is 0.18 support.
#wlroots is great to wrap up the so-called boilerplate, but it's an absolute joke for developers trying to hop from release-to-release. The documentation sucks, the rationale is non-existent for changes made, and I've had to build each 0.X release to then `git-diff` between the releases.
wlroots needs to do a lot better at not fucking with the API so much -- and if they must -- they must do better to document where the changes are breaking, why, and what the expected fix is.
I shouldn't have to `git diff` releases to determine this.


