- Wednesday: I decided to devote more time to working on MOTR. I took an inventory of the flake8 packages I'm using, and tried to move flags into config for a more reliable experience.
- Thursday: I thought I was ready to write tests, until I realized that there isn't a good interface to access all of this code from outside of MOTR.
- Friday: I took a break from MOTR to mess around with BQN.
- Saturday: I messed a little more with BQN, and took an inventory of some of MOTR's code to see which modules I want to move around.
- Sunday: I reorganized and exposed all of the command-line wrappers for MOTR, including the ones I haven't written yet, which is most of them.
- Monday: I put together type testing as part of an effort to actually use the "registry" code I wrote for the higher-level interface, because that code is the only non-trivial module that my tests don't exercise at all. After writing some tests, I concluded that the plugin code required to make this all work is sufficiently weird that this is a sign that I'm pushing Mypy too far and I should back off, which is a decision that I definitely held to today.
Next week, I'm going to see about having a good balance between MOTR and not-MOTR, but no promises.