mirror of https://github.com/astral-sh/ruff
50f14d017e
## Summary This PR updates the linter benchmark to use the `tokenize` function instead of the lexer. The linter expects the token list to be up to and including the first error which is what the `ruff_python_parser::tokenize` function returns. This was not a problem before because the benchmarks only uses valid Python code. |
||
|---|---|---|
| .. | ||
| benches | ||
| src | ||
| Cargo.toml | ||
| README.md | ||
README.md
Ruff Benchmarks
The ruff_benchmark crate benchmarks the linter and the formatter on individual files:
# Run once on the "baseline".
cargo bench -p ruff_benchmark -- --save-baseline=main
# Compare against the "baseline".
cargo bench -p ruff_benchmark -- --baseline=main
# Run the lexer benchmarks.
cargo bench -p ruff_benchmark lexer -- --baseline=main
See CONTRIBUTING.md on how to use these benchmarks.