Three Dollar Quill

• Wednesday: Three tests ported over, and I managed to simultaneously lament my hubris and continue to engage in it.
• Thursday: Four tests ported over. A little worried about trying to port over the random number stuff, since C# and Python provide different random generators out of the box.
• Friday: Six tests. Getting kind of tired of all of the bits where I mixed up "and" and "or" or "parent" and "first child". I think I also managed to push some values to the wrong stack, which took a surpisingly long time to mess things up.
• Saturday: More dealing with hubris.
• Sunday: I did a quick look at the syntactic bestiary, then tried to do a semantic analysis of part of the post I was writing. Even at the time, it was quickly clear to me that it was too late at night for me to be trying that.
• Monday: A bit more unfocused work.

Next week, definitely more porting Ink. This is an interesting experience. While I do have specific complaints about parts of Ink's code, overall I feel the codebase has earned far more respect from me than I had for the code from TStand90's roguelike tutorials. It's a matter of, the requirements put on Ink are much stricter than those on the tutorial code. The ultimate form of the VM will probably actually end up extremely similar to the C# code. I'd like to pull off some kind of bonkers refactor, for the dopamine hit, but the likely targets for improvement are the awful interfaces I've imposed on it. (Every three tests or so, I'm just like "I've made a huge mistake, but I can cope".)

On the conlanging side, I want to clean up my glosses and make them cross-referenceable somehow. I've got some ideas I'll look into for that.