whitehole

A simple, fast, intuitive parser combinator framework for Rust.

Features

• Simple: only a handful of combinators to remember: eat, next, till, wrap.
• Operator overloading: use + and | to compose combinators, use * to repeat a combinator.
• Zero heap allocation: this framework only uses stack memory.
• Re-usable heap memory: if you need allocation, never clone or reallocate it.
• Stateful-able: control the parsing flow with an optional custom state.
• Safe by default, with unsafe variants for performance.

Installation

cargo add whitehole

Examples

Here is a simple example to parse hexadecimal color codes:

use whitehole::{
  combinator::{eat, next},
  parser::Builder,
};

let double_hex = || {
  // Repeat a combinator with `*`.
  (next(|c| c.is_ascii_hexdigit()) * 2)
    // Convert the matched content to `u8`.
    .select(|ctx| u8::from_str_radix(ctx.content(), 16).unwrap())
    // Wrap `u8` to `(u8,)`, this is required by `+` below.
    .tuple()
};

// Concat multiple combinators with `+`.
// Tuple values will be concatenated into a single tuple.
// Here `() + (u8,) + (u8,) + (u8,)` will be `(u8, u8, u8)`.
let entry = eat('#') + double_hex() + double_hex() + double_hex();

let mut parser = Builder::new().entry(entry).build("#FFA500");
let output = parser.parse().unwrap();
assert_eq!(output.digested, 7);
assert_eq!(output.value, (255, 165, 0));

Documentation

Related

• in_str: a procedural macro to generate a closure that checks if a character is in the provided literal string.

Credits

This project is inspired by:

• nom
• combine
• tree-sitter
• retsac

CHANGELOG