Following along the first (Java) part of https://craftinginterpreters.com/, but with Rust, because reasons.
Status: done!
Check out the interpreter in action in your browser at https://abesto.github.io/jlox-rs/ (and click "What am I looking at?" for tons of details)
About half as fast as the canonical jlox implementation. Considering that's Java and this is Rust, this is somewhat surprising. Considering this is the first time I ever build an interpreter, and did zero optimization, this is very surprising, but in the opposite direction.
Benchmark script:
fun fib(n) {
if (n < 2) return n;
return fib(n - 1) + fib(n - 2);
}
print fib(30);
Benchmark results:
$ hyperfine --warmup 1 '../craftinginterpreters/jlox ./fib.lox' './target/release/jlox_rs ./fib.lox'
Benchmark 1: ../craftinginterpreters/jlox ./fib.lox
Time (mean ± σ): 855.8 ms ± 196.5 ms [User: 1134.4 ms, System: 106.2 ms]
Range (min … max): 695.4 ms … 1151.9 ms 10 runs
Benchmark 2: ./target/release/jlox_rs ./fib.lox
Time (mean ± σ): 2.016 s ± 0.026 s [User: 2.011 s, System: 0.002 s]
Range (min … max): 1.982 s … 2.057 s 10 runs
Summary
'../craftinginterpreters/jlox ./fib.lox' ran
2.36 ± 0.54 times faster than './target/release/jlox_rs ./fib.lox'
I added my own tests as I went, and then at the very end found that https://github.com/munificent/craftinginterpreters contains a full test suite. Running it against this implementation, I see 100 passes and 139 failures. A ton of the failures are due to error message formatting. A handful might not be, but honestly that's fine for this hobby project. I'll probably use those tests for clox-rs though!