pre-commit by Yelp

A framework for managing and maintaining multi-language pre-commit hooks.

Build Status Coverage Status

At Yelp we rely heavily on pre-commit hooks to find and fix common issues before changes are submitted for code review. We run our hooks before every commit to automatically point out issues like missing semicolons, whitespace problems, and testing statements in code. Automatically fixing these issues before posting code reviews allows our code reviewer to pay attention to the architecture of a change and not worry about trivial errors.

As we created more libraries and projects we recognized that sharing our pre commit hooks across projects is painful. We copied and pasted bash scripts from project to project and had to manually change the hooks to work for different project structures.

We believe that you should always use the best industry standard linters. Some of the best linters are written in languages that you do not use in your project or have installed on your machine. For example scss-lint is a linter for SCSS written in Ruby. If you’re writing a project in node you should be able to use scss-lint as a pre-commit hook without adding a Gemfile to your project or understanding how to get scss-lint installed.

We built pre-commit to solve our hook issues. It is a multi-language package manager for pre-commit hooks. You specify a list of hooks you want and pre-commit manages the installation and execution of any hook written in any language before every commit. pre-commit is specifically designed to not require root access. If one of your developers doesn’t have node installed but modifies a JavaScript file, pre-commit automatically handles downloading and building node to run jshint without root.

Before you can run hooks, you need to have the pre-commit package manager installed.

Using pip:

pip install pre-commit

Non Administrative Installation:

curl http://pre-commit.com/install-local.py | python

System Level Install:

curl https://bootstrap.pypa.io/get-pip.py | sudo python - pre-commit

In a Python Project, add the following to your requirements.txt (or requirements-dev.txt):

pre-commit

Once you have pre-commit installed, adding pre-commit plugins to your project is done with the .pre-commit-config.yaml configuration file.

Add a file called .pre-commit-config.yaml to the root of your project. The pre-commit config file describes:

repo, sha where to get plugins (git repos). sha can also be a tag.
id What plugins from the repo you want to use.
language_version (optional) Override the default language version for the hook. See Advanced Features: "Overriding Language Version".
files (optional) Override the default pattern for files to run on.
exclude (optional) File exclude pattern.
args (optional) additional parameters to pass to the hook.

For example:

-   repo: git://github.com/pre-commit/pre-commit-hooks
    sha: 5541a6a046b7a0feab73a21612ab5d94a6d3f6f0
    hooks:
    -   id: trailing-whitespace

This configuration says to download the pre-commit-hooks project and run its trailing-whitespace hook.

Run pre-commit install to install pre-commit into your git hooks. pre-commit will now run on every commit. Every time you clone a project using pre-commit running pre-commit install should always be the first thing you do.

If you want to manually run all pre-commit hooks on a repository, run pre-commit run --all-files. To run individual hooks use pre-commit run <hook_id>.

The first time pre-commit runs on a file it will automatically download, install, and run the hook. Note that running a hook for the first time may be slow. For example: If the machine does not have node installed, pre-commit will download and build a copy of node.

pre-commit currently supports hooks written in JavaScript (node), Python, Ruby and system installed scripts. As long as your git repo is an installable package (gem, npm, pypi, etc.) or exposes an executable, it can be used with pre-commit. Each git repo can support as many languages/hooks as you want.

An executable must satisfy the following things:

  • Returncode of hook must be different between success / failures (Usually 0 for success, nonzero for failure)
  • It must take filenames

A git repo containing pre-commit plugins must contain a hooks.yaml file that tells pre-commit:

id The id of the hook - used in pre-commit-config.yaml
name The name of the hook - shown during hook execution
entry The entry point - The executable to run
files The pattern of files to run on.
language The language of the hook - tells pre-commit how to install the hook.
description (optional) The description of the hook.
language_version (optional) See Advanced Features: "Overriding Language Version".
expected_return_value (optional) Defaults to 0.

For example:

-   id: trailing-whitespace
    name: Trim Trailing Whitespace
    description: This hook trims trailing whitespace.
    entry: trailing-whitespace-fixer
    language: python
    files: \.(js|rb|md|py|sh|txt|yaml|yml)$

Supported languages

  • node
  • python
  • ruby
  • pcre - "Perl Compatible Regular Expression" Specify the regex as the entry
  • script - A script existing inside of a repository
  • system - Executables available at the system level

Running in migration mode

By default, if you have existing hooks pre-commit install will install in a migration mode which runs both your existing hooks and hooks for pre-commit. To disable this behavior, simply pass -f / --overwrite to the install command. If you decide not to use pre-commit, pre-commit uninstall will restore your hooks to the state prior to installation.

Temporarily disabling hooks

Not all hooks are perfect so sometimes you may need to skip execution of one or more hooks. pre-commit solves this by querying a SKIP environment variable. The SKIP environment variable is a comma separated list of hook ids. This allows you to skip a single hook instead of --no-verifying the entire commit.

$ SKIP=flake8 git commit -m "foo"

pre-commit during commits

Running hooks on unstaged changes can lead to both false-positives and false-negatives during committing. pre-commit only runs on the staged contents of files by temporarily saving the contents of your files at commit time and stashing the unstaged changes while running hooks.

pre-commit during merges

The biggest gripe we’ve had in the past with pre-commit hooks was during merge conflict resolution. When working on very large projects a merge often results in hundreds of committed files. I shouldn’t need to run hooks on all of these files that I didn’t even touch! This often led to running commit with --no-verify and allowed introduction of real bugs that hooks could have caught. pre-commit solves this by only running hooks on files that conflict or were manually edited during conflict resolution.

pre-commit during push

As of version 0.3.5, pre-commit can be used to manage pre-push hooks. Simply pre-commit install --hook-type pre-push.

Passing arguments to hooks

Sometimes hooks require arguments to run correctly. You can pass static arguments by specifying the args property in your .pre-commit-config.yaml as follows:

-   repo: git://github.com/pre-commit/pre-commit-hooks
    sha: 5541a6a046b7a0feab73a21612ab5d94a6d3f6f0
    hooks:
    -   id: flake8
        args: [--max-line-length=131]

This will pass --max-line-length=131 to flake8.

Arguments Pattern in hooks

If you are writing your own custom hook as a script-type or even a system hook, your hook should expect to receive the args value and then a list of staged files.

For example, assuming your .pre-commit-config.yaml was like below

-   repo: git://github.com/path/to/your/hook/repo
    sha: 5541a6a046b7a0feab73a21612ab5d94a6d3f6f0
    hooks:
    -   id: my-hook-script-id
        args: [--myarg1=1 --myarg1=2]

When you next run pre-commit, it will pass on those args as follows:

path/to/script-or-system-call --myarg1=1 --myarg1=2 dir/file1 dir/file2 file3

If the args property is empty or not defined, you should expect your script to be called:

 path/to/script-or-system-call dir/file1 dir/file2 file3 

Overriding Language Version

Sometimes you only want to run the hooks on a specific version of the language. For each language, they default to using the system installed language (So for example if I’m running python2.6 and a hook specifies python, pre-commit will run the hook using python2.6). Sometimes you don’t want the default system installed version so you can override this on a per-hook basis by setting the language_version.

-   repo: git://github.com/pre-commit/mirrors-scss-lint
    sha: d7266131da322d6d76a18d6a3659f21025d9ea11
    hooks:
    -   id: scss-lint
        language_version: 1.9.3-p484

This tells pre-commit to use 1.9.3-p484 to run the scss-lint hook.

Valid values for specific languages are listed below:

  • python: Whatever system installed python interpreters you have. The value of this argument is passed as the -p to virtualenv.
  • node: See nodeenv.
  • ruby: See ruby-build

Usage in Continuous Integration

pre-commit can also be used as a tool for continuous integration. For instance, adding pre-commit run --all-files as a CI step will ensure everything stays in tip-top shape.

We’re looking to grow the project and get more contributors especially to support more languages/versions. We’d also like to get the hooks.yaml files added to popular linters without maintaining forks / mirrors.

Feel free to submit Bug Reports, Pull Requests, and Feature Requests.

When submitting a pull request, please enable travis-ci for your fork.