I did a little writing today, but mostly I translated the tox template into a noxfile. This took all day to get right, and required careful finesse to get working on my machine. I'm currently doing a mutation run to see how the changes I made shook out. Current impressions: not great, but maybe I can make some improvements.
I got bored with how long this is taking, so I went and made the improvements. I can't tell if it did anything, but it really should have. Less IO, fewer virtual environments, fewer sessions. Probably the right way to go with this would be to take this one hot path, and write a shell script for it or something. Maybe have a nox session just for setting up the test environment. Maybe throw together a tiny wrapper in Rust because I don't want to figure out how to do string processing in bash.
These are interesting ideas that I am currently too sleepy to do justice to. I can at least get things set up to develop for it, though.
I say that, and then I go and get a project set up that has all of the high-level functions needed, and would compile, but wouldn't actually do anything, because I'm too sleepy to implement the helper functions.
There's nothing more to be done now.