kuroko-wasm-repl VS rlox

Kuroko 1.3.0 gets a release candidate. Lots of big things since 1.2.5, like optimized method invocation, more operator overloads, better support for f-string expressions (format specs, =, faster constructions), a long type with my own bigint implementation (this was the last thing I was still regularly opening Python repls for, so a huge personal win). I also fixed a bunch of little things that have been nagging me, like the compiler can now compile expressions directly, which allowed me to remove the kludge that made the repl work previously. The WASM web repl also got some love with a port of the core of Hiwire from Pyodide, giving a very straightforward interface between JS and Kuroko in a browser - and I rebuilt the web IDE on it. I've also been working on a new compiler, which will hopefully form the basis of 2.0 - and this might be the last 1.x release (though I expect at least a few 1.3.x bug fix releases).

It's actually the other way round: part one implements an AST walking interpreter in Java while the second part builds an interpreter in C with compilation to byte code for a virtual machine that is also developed from scratch. Currently I try to build the second part solution in Rust (https://github.com/thomasbollmeier/rlox).

Great, I do the same:-). I have however only proceeded with my implementation to the Pratt expression parsing so far: https://github.com/thomasbollmeier/rlox

I have just started with a rust implementation of the VM and the interpreter from part 2 (https://github.com/ThomasBollmeier/rlox). While following the book I wonder if it wouldn't be easier to build on the interpreter from part 1 and just emit bytecode while traversing the AST.