tree-sitter-qsharp

A tree-sitter grammar for Q#, Microsoft's quantum programming language.

This is a new project, not affiliated with Microsoft or the tree-sitter organization. Feedback, bug reports, and contributions are welcome.

What it does

Parses Q# source files into a concrete syntax tree that editors can use for syntax highlighting, code folding, symbol navigation, and indentation.

Built by reading the Q# language reference and the Q# compiler source. Tested against Q# files in the microsoft/qsharp repository.

Supported constructs

• Namespaces, imports, exports, open directives
• Operations and functions with type parameters, functor clauses, specializations
• Structs, newtypes, all builtin types
• Full expression grammar with operator precedence per the official specification
• Quantum-specific syntax: use/borrow, within/apply, repeat/until/fixup, Adjoint/Controlled
• All literals: int, bigint, float, imaginary, string, interpolated string, bool, Result, Pauli
• External scanner for float/range disambiguation (1. vs 0..5)

Known limitations

• Explicit type arguments on calls (Foo<Int>(x)) are not supported due to </> ambiguity with comparison operators. Q# type inference handles this in practice.
• Open-ended step ranges (...2...) parse but the AST does not perfectly represent all range component positions.
• This is a v0.x project. There may be edge cases that are not yet covered.

Editor query files

Capture names follow the nvim-treesitter conventions.

Contributing

If you find Q# code that doesn't parse correctly, please open an issue with the .qs file or a minimal reproducing snippet. Pull requests are welcome.

References

• Q# overview
• microsoft/qsharp — Q# compiler and standard library (MIT)
• tree-sitter — parser generator

License

MIT