Three Dollar Quill

Aside from some gaps in mutation coverage, I believe the code is in a good, if transitional, state. Basically, I got the tests to a good state (aside from some gaps in mutation coverage), and then I refactored things so that there is no behavior on syntax nodes, and instead all the behavior lives in singledispatch functions. The goal here is to now parameterize the node types, and then only have the singledispatch functions be defined for specific type instances.

My hope is that I can then reuse individual function implementations by making the signatures just permissive enough that I can register them in multiple places.

If this works, it'll be great. If it doesn't, then I just kind of exploded my code for no real benefit.

You'll have to wait and see, because I'm definitely not trying anything else tonight.

Good night.