clang-tidy 22.1.8

Clang-tidy is an LLVM-based code analyser tool

clang-tidy Python distribution

This project packages the clang-tidy utility and related tools as a Python package. It allows you to install clang-tidy directly from PyPI:

python -m pip install clang-tidy

The tools provided are:

• clang-tidy: lints C++ source code using configurable rules
• clang-apply-replacements: applies "fixes" that have been exported by clang-tidy in YAML format
• run-clang-tidy.py: runs clang-tidy over all files in a compilation database or a specified path, with parallelism
• clang-tidy-diff.py: applies clang-tidy to the changed lines in a user-provided diff

This projects intends to release a new PyPI package for each major and minor release of clang-tidy.

Use with pipx

You can use pipx to run clang-tidy, as well. For example, pipx run clang-tidy <args> will run clang-tidy without any previous install required on any machine with pipx (including all default GitHub Actions / Azure runners, avoiding requiring a pre-install step or even actions/setup-python).

Building new releases

The clang-tidy-wheel repository provides the logic to build and publish binary wheels of the clang-tidy utility.

In order to add a new release, the following steps are necessary:

• Edit the version file to reflect the new version.
• Make a GitHub release to trigger the GitHub Actions release workflow. Alternatively, the workflow can be triggered manually.

On manual triggers, the following input variables are available:

• use_qemu: Whether to build targets that require emulation (default: true)
• llvm_version: Override the LLVM version (default: "")
• wheel_version: Override the wheel packaging version (default "0")
• deploy_to_testpypi: Whether to deploy to TestPyPI instead of PyPI (default: false)

The repository with the precommit hook is automatically updated using a scheduled Github Actions workflow.

Acknowledgments

This repository extends the great work of several other projects:

• clang-tidy itself is provided by the LLVM project under the Apache 2.0 License with LLVM exceptions.
• The build logic is based on scikit-build which greatly reduces the amount of low level code necessary to package clang-tidy.
• The scikit-build packaging examples of CMake and Ninja were very helpful in packaging clang-tidy.
• The CI build process is controlled by cibuildwheel which makes building wheels across a number of platforms a pleasant experience (!)

We are grateful for the generous provisioning with CI resources that GitHub currently offers to Open Source projects.

Troubleshooting

To see which clang-tidy binary the package is using you can set CLANG_TIDY_WHEEL_VERBOSE to 1 in your environment.