Overview

LogoSky is a Rust library that bridges Logos and Chumsky, combining the speed of Logos' lexical analysis with the expressiveness of Chumsky's parser combinators. It provides a seamless integration layer that allows you to use Logos for tokenization and Chumsky for parsing in a unified, type-safe manner.

Features

• 🚀 High-performance tokenization with Logos
• 🎯 Expressive parser combinators with Chumsky
• 🔗 Seamless integration via TokenStream adapter
• 📍 Rich span tracking with Span and Spanned types
• 🎨 Flexible parseable trait for building composable parsers
• 🔧 No-std support with optional allocator
• 🧩 Multiple source types support (str, [u8], Bytes, BStr, HipStr)
• ⚡ Zero-cost abstractions for efficient parsing

Installation

Add this to your Cargo.toml:

[dependencies]
logosky = "0.2"

Feature Flags

• std (default) - Enable standard library support
• alloc - Enable allocator support for no-std environments
• bytes - Support for bytes::Bytes as token source
• bstr - Support for bstr::BStr as token source
• hipstr - Support for hipstr::HipStr as token source
• either - Enable Either<L, R> parseable support
• among - Enable Among<L, M, R> parseable support
• smallvec - Enable small vector optimization utilities

Core Components

• TokenStream<'a, T>

  An zero-copy adapter that bridges Logos lexer output to Chumsky parser input. Implements Chumsky's Input, ValueInput, SliceInput, and ExactSizeInput traits.

• Token<'a> Trait

  The core trait for defining tokens. Requires:

  • Char: Character type (usually char or u8)
  • Kind: Token kind discriminator
  • Logos: Associated Logos token enum

• LosslessToken<'a> Trait

  Extends Token<'a> for tokens that preserve all source information, including trivia (whitespace, comments).

  • Provides is_trivia() method to identify non-semantic tokens
  • Essential for building formatters, linters, and language servers that need to preserve formatting
  • Works seamlessly with Tokenizer trivia handling utilities

• Tokenizer<'a, T> Trait

  Provides parser combinators for working with token streams:

  • skip_trivias(): Skip over trivia tokens (whitespace, comments)
  • collect_trivias(): Collect trivia tokens into a container for later processing

• Parseable<'a, I, T, Error> Trait

  A trait for types that can be parsed from a token stream. Implement this to create composable parsers.

• Span and Spanned<T>

  • Span: Lightweight span tracking (start/end positions)
  • Spanned<T>: Wraps a value with its source span

• Utility Traits

  • IsAsciiChar: ASCII character checking utilities
  • AsSpan / IntoSpan: Span access traits
  • IntoComponents: Destructure parsed elements

Who uses logosky?

• smear: Blazing fast, fully spec-compliant, reusable parser combinators for standard GraphQL and GraphQL-like DSLs.

License

logosky is dual-licensed under:

• MIT License (LICENSE-MIT)
• Apache License, Version 2.0 (LICENSE-APACHE)

You may choose either license for your purposes.

Copyright (c) 2025 Al Liu.