Three Dollar Quill

• Wednesday: I made some minor adjustments to the Ink port code. I broke things, and decided not to delay the entry trying to fix them.
• Thursday: I fixed the issue that was entirely my fault, then started trying to work out what I actually need to track as part of the state object. This is tricky because the member declarations seem to be scattered throughout the class bodies, which in some of these files are nested, and also I'm not used to languages with implicit self. Also, the whole "we need a zero-argument constructor for deserialization" thing adds a whole extra layer of indirection. (My code does not explicitly attempt to have zero-argument constructors.)
• Friday: Putting that stuff into practice was helpful, because I'd messed up the layout of LineState, because it's hard to guess whether something is a member or a getter/setter when I don't have a clear idea of where to look for the definitions.
• Saturday: I cut over to the new layout, and got things exactly as broken as before.
• Sunday: A bit of work conlanging, focusing on the broad contours of what is and isn't in the syntax.
• Monday: A little bit more of that, but not much.

Next week, I'm really getting into this Ink port stuff, so I see no reason to stop right now. I'm up to two out of three ported tests passing; I'm kind of hoping to get some kind of Pareto situation where most of the effort is in getting the basics to work, and things are just minor tweaks by the time I get to the hundredth test.