I ended up just adapting a single project to flit today. I pulled in a bunch of stuff from Structured Data, and discovered that my whole hack to get around pip install . doesn't actually work on flit (yet, there's a PR), but the install is fast enough that it turns out not to matter in practice, at least for this project. Now that I've got everything together nicely for this project, I just need to write tests for the code, and then I can start working on the various refactors I have in mind.
I think the next repository I'll look into adapting is from my latest attempt to port/rewrite Ink. In some ways that'll be a little tricky, because it's not that there are no tests, it's that the tests don't pass because I haven't ported enough logic over yet. Oh well.
At some point I'll want to try out mutation testing against something, but that requires tests, that work. My best candidate is probably Structured Data.
Once I've done as much with flit as possible (probably up to porting the Cell nucleus file for Python projects to use it), I'll try to learn SCons. I wouldn't terribly mind putting that off.
Speaking of putting things off, I need to post. Now.
(Last note, I feel like mentioning that my PR to Poetry finally got an approval shortly after the last post got published. The merging is still blocked, which I don't feel like figuring out right now.)