UPDATE #25

[ityonemo]

Happy New Year!

After taking a bit longer than I expected updating zigler, I am back in the saddle on working on Selectrix. Over in mavis_inference land, I ran into a bit of snag because some test opcode statements are not working correctly, and I made a video about the surprising meaning of is_nil in the virtual machine: https://youtu.be/nebDOqU0TJ0

I decided to go ahead and work on some stuff in mavis that will help out with test opcodes, and that is LITERAL TYPES. This has been nagging at me ever since I started this project, and quite frankly it's disappointing that dialyzer does not support them. PR to merge Literal Type functionality is here: https://github.com/ityonemo/mavis/pull/140 and I will be making a video on the subject soon. It was also really neat how smoothly literal types fit in the organizational scheme of Mavis, which to me is a sign that "I am doing things correctly".

Now that Literal types are done, it's going to be much easier writing code for SUBTRACTION TYPES. This is explicitly not a part of what Dialyzer supports (nor does Gradualizer support this). It's going to be a bit of effort, but I think that the way that mavis is organized is going to make producing Subtraction types easy. And even though it's going to make Mavis more complex, it's going to be worth it because test conditions, and case statements, etc. are going to be buttery-smooth.

Update 24: https://github.com/ityonemo/selectrix/issues/33