parsejevko.js VS nederlang

Besides that I wrote some parsers in various languages. The most mature is the JS one: https://github.com/jevko/parsejevko.js

Writing a blog post on my experience rewriting a toy interpreter in Rust and what it took to make it faster than the one I wrote in C.

Closures will probably remain hairy though. It’s a tough problem to solve. I’ve yet to take the plunge (supporting closures, that is) for my toy language Nederlang but maybe the other parts of it can be of help. It’s no longer a tree-walking implementation though.

The way I solved this in Nederlang is to create a hole in the stack for the maximum amount of local variables the compiler encountered for each block statement.

Have been working on Nederlang for the last few weeks again. My last (and first!) attempt at an interpreted programming language was in C but it wasn't that pleasant to work in. I'm now using Rust and am pretty happy with it so far. Performance is great (using tagged pointers and some nice optimizations using specialized opcodes for common operations, like operating on a constant and a variable, bypassing the stack), it's a joy to work in and exotic segfaults are a thing of the past.