tree-sitter-solidity

tree-sitter-solidity is a Tree-sitter grammar for Solidity, the smart-contract language for Ethereum and EVM chains. It parses contract source into a fast, incrementally-updated concrete syntax tree that editors and analysis tools can walk for highlighting, navigation, and metaprogramming.

The grammar aims to be an efficient, low-dependency parser that targets most Solidity versions in use, and ships bindings for Rust, C, Node, and other Tree-sitter hosts from a single definition.

What You Get

• A Tree-sitter grammar for Solidity targeting most versions in use, compiled to a portable C parser.
• A Rust binding exposing the parser as a tree_sitter_language::LanguageFn for the Tree-sitter runtime.
• Highlight and injection query files ready for editor integration.
• A low-dependency parse tree suitable for metaprogramming and static analysis.

Common Use Cases

• Syntax highlighting for .sol files in editors like Neovim.
• Structural navigation, folding, and text objects over Solidity contracts.
• Static analysis, linting, and metaprogramming tools that need a Solidity parse tree.

Under The Hood

Architecture - The grammar is authored in grammar.js using Tree-sitter’s DSL and compiled to a portable src/parser.c plus src/node-types.json; language bindings each wrap that generated parser, and query files provide editor highlighting. The grammar draws structure and inspiration from tree-sitter-javascript and Ethereum’s official Solidity ANTLR grammar.

Tech Stack - The Rust binding (built via build.rs with the cc crate) depends on tree-sitter-language at runtime and tree-sitter for testing; the repo also carries npm and other manifests so one grammar serves multiple ecosystems, with Node.js CI validating changes.

Code Quality - Correctness is anchored by an extensive Tree-sitter test corpus under test/, exercised through the standard tree-sitter test harness; the README notes the generated AST structure is still stabilizing, so tests are the primary contract.

API Design - The public surface follows Tree-sitter conventions: the crate exports a LANGUAGE constant (LanguageFn) and query strings, so anyone familiar with the ecosystem can wire it into a parser in a few lines.