I got up to 27 tests ported over the weekend, though two of them kind of freak me out; it's not clear to me how they pass in the C# version, but I verified for myself that they do, so this is my problem.
Anyway, I know the next tests to go for will work out poorly for me, because I've been really lax in populating the error/warning data. That's why I spent today shoring up the typing system and discovering a bug thanks to that. I'm going to try to psych myself up to fix the error/warning code tomorrow, so I can hopefully keep breezing through tests after that.
I also realized yesterday that if I relax some preconceptions I had about how it should all work, then Sphinx might be a good fit for conlanging work. Lets me link stuff from my glosses, adding a lexicon entry is relatively lightweight, and if I get really desperate for some kind of generated data, I can do terrible things to the contents of Python docstrings.
I'm going to work on loading my disparate notes into it, when I'm not up for puzzling out what's happening in Ink. (I'm honestly still not sure what's actually different between how I had some code implemented, and what I changd it to; I just know that the former was an attempt to match my understanding of Ink's continue logic, and the latter is an extremely close translation of the actual code. They do not behave the same.)
Anyway, that's enough for now. Maybe I'll actually get to bed in a timely fashion tonight.