Three Dollar Quill

• Wednesday: I got back into the hobby project that originally prompted me to write the pip PR, and planned some short-term improvements.
• Thursday: I got a refactor partway done by the time I published, and I think I finished it later that night.
• Friday: I decided that "my laptop is stupid beefy, my code evaluation tasks are somewhat embarrassingly parallel, and the task runner I'm currently using has no concurrency" was a bottleneck for iteration. I wrote up the feature set that I wanted, and listed a bunch of tools that didn't satisfy it. I haven't heard back any suggestions, so shortly afterwards, I decided to try rolling my own, even as I suspected that doing so would open me up to all sorts of weird complications. Truly tragic.
• Saturday: I did various bits of housekeeping, of varying levels of consequence.
• Sunday: I prototyped the "fine, I'll do it myself" task runner, and researched the few bits of my current task definitions that were potentially contending for file access.
• Monday: I started getting a little more serious about the task runner, and trying to nail down what it's behavior should be.

Next week, I don't quite want to touch code for any of my projects, so I'm going to keep up with planning and prototyping the runner, and doing stuff not related to code.