Revisit `%` operator semantics

[osa1]

Currently % works the same way as Rust %.

This is not a deliberate choice: I needed a remainder operation with positive dividend and divisor. And because the interpreter is in Rust, using Rust's % for this operation was the easiest implementation that gave me what I wanted.

For the long term though, we should consider the alternative implementations. I found this blog post which has a survey of different implementations in different languages, and thoughts on which one should be used in new languages: https://julesjacobs.com/2023/09/23/on-div-and-mod.html.