pre-commit

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

Build Status Coverage Status

Git hook scripts are useful for identifying simple issues before submission to code review. We run our hooks on every commit to automatically point out issues in code such as missing semicolons, trailing whitespace, and debug statements. By pointing these issues out before code review, this allows a code reviewer to focus on the architecture of a change while not wasting time with trivial style nitpicks.

As we created more libraries and projects we recognized that sharing our pre-commit hooks across projects is painful. We copied and pasted unwieldy 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 eslint 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: (To upgrade: run again, to uninstall: pass uninstall to python)

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

Using homebrew

brew install 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.
types (optional) Override the default file types to run on. See Filtering files with types. new in 0.15.0
exclude_types (optional) File types exclude. new in 0.15.0
args (optional) additional parameters to pass to the hook.
stages (optional) Confines the hook to the commit, push, or commit-msg stage. See Advanced Features: "Confining Hooks To Run At A Certain Stage".
additional_dependencies (optional) A list of dependencies that will be installed in the environment where this hook gets run. One useful application is to install plugins for hooks such as eslint. new in 0.6.6
always_run (optional) Default false. If true this hook will run even if there are no matching files. new in 0.7.2

For example:

-   repo: git://github.com/pre-commit/pre-commit-hooks
    sha: v0.9.1
    hooks:
    -   id: trailing-whitespace

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

Updating hooks automatically

You can update your hooks to the latest version automatically by running pre-commit autoupdate. This will bring the hooks to the latest sha on the master branch.

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.

The hook must exit nonzero on failure or modify files in the working directory (since 0.6.3).

A git repo containing pre-commit plugins must contain a .pre-commit-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
language The language of the hook - tells pre-commit how to install the hook.
files (optional) Default ''. The pattern of files to run on. new in 0.15.0: now optional
exclude (optional) Default ^$. Exclude files that were matched by files.
types (optional) Default [file]. List of file types to run on. See Filtering files with types. new in 0.15.0
exclude_types (optional) Default []. Exclude files that were matched by types. new in 0.15.0
always_run (optional) Default false. If true this hook will run even if there are no matching files. new in 0.7.2
pass_filenames (optional) Default true. If true this hook must take filenames as positional arguments. new in 0.14.0
description (optional) The description of the hook.
language_version (optional) See Advanced Features: "Overriding Language Version".
minimum_pre_commit_version (optional) Allows one to indicate a minimum compatible pre-commit version. new in 0.6.7

For example:

-   id: trailing-whitespace
    name: Trim Trailing Whitespace
    description: This hook trims trailing whitespace.
    entry: trailing-whitespace-fixer
    language: python
    types: [text]

new in 0.12.0 Prior to 0.12.0 the file was hooks.yaml (now .pre-commit-hooks.yaml). For backwards compatibility it is suggested to provide both files or suggest users use pre-commit>=0.12.0.

Supported languages

docker

new in 0.10.0

The hook repository must have a Dockerfile. It will be installed via docker build ..

Running Docker hooks requires a running Docker engine on your host. For configuring Docker hooks, your entry should correspond to an executable inside the Docker container, and will be used to override the default container entrypoint. Your Docker CMD will not run when pre-commit passes a file list as arguments to the run container command. Docker allows you to use any language that's not supported by pre-commit as a builtin.

Support: docker hooks are known to work on any system which has a working docker executable. It has been tested on linux and macOS. Hooks that are run via boot2docker are known to be unable to make modifications to files.

See this repository for an example Docker-based hook.

golang

new in 0.12.0

The hook repository must contain go source code. It will be installed via go get ./.... pre-commit will create an isolated GOPATH for each hook and the entry should match an executable which will get installed into the GOPATH's bin directory.

Support: golang hooks are known to work on any system which has go installed. It has been tested on linux, macOS, and windows.

node

The hook repository must have a package.json. It will be installed via npm install .. The installed package will provide an executable that will match the entry – usually through bin in package.json.

Support: node hooks work without any system-level dependencies. It has been tested on linux and macOS and may work under cygwin.

python

The hook repository must have a setup.py. It will be installed via pip install .. The installed package will provide an executable that will match the entry – usually through console_scripts or scripts in setup.py.

Support: python hooks work without any system-level depedendencies. It has been tested on linux, macOS, windows, and cygwin.

ruby

The hook repository must have a *.gemspec. It will be installed via gem build *.gemspec && gem install *.gem. The installed package will produce an executable that will match the entry – usually through exectuables in your gemspec.

Support: ruby hooks work without any system-level dependencies. It has been tested on linux and macOS and may work under cygwin.

swift

new in 0.11.0

The hook repository must have a Package.swift. It will be installed via swift build -c release. The entry should match an executable created by building the repository.

Support: swift hooks are known to work on any system which has swift installed. It has been tested on linux and macOS.

pcre

"Perl Compatible Regular Expressions" — pcre hooks are a quick way to write a simple hook which prevents commits by file matching. Specify the regex as the entry.

osx does not ship with a functioning grep -P so you'll need brew install grep for pcre hooks to function.

Support: pcre hooks work on any system which has a functioning grep -P (or in the case of macOS: ggrep -P). It has been tested on linux, macOS, windows, and cygwin.

script

Script hooks provide a way to write simple scripts which validate files. The entry should be a path relative to the root of the hook repository.

This hook type will not be given a virtual environment to work with – if it needs additional dependencies the consumer must install them manually.

Support: the support of script hooks depend on the scripts themselves.

system

System hooks provide a way to write hooks for system-level executables which don't have a supported language above (or have special environment requirements that don't allow them to run in isolation such as pylint).

This hook type will not be given a virtual environment to work with – if it needs additional dependencies the consumer must install them manually.

Support: the support of system hooks depend on the executables.

Developing hooks interactively

Since the repo property of .pre-commit-config.yaml can take anything that git clone ... understands, it's often useful to point it at a local directory on your machine while developing hooks and using pre-commit autoupdate to synchronize changes.

-   repo: /home/asottile/workspace/pre-commit-hooks
    sha: v0.9.1
    hooks:
    -   id: trailing-whitespace

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. This also includes files which were automatically merged by git. Git isn't perfect and this can often catch implicit conflicts (such as with removed python imports).

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.

pre-commit for commit messages

new in 0.15.4 pre-commit can be used to manage commit-msg hooks. Simply pre-commit install --hook-type commit-msg.

commit-msg hooks can be configured by setting stages: [commit-msg]. commit-msg hooks will be passed a single filename -- this file contains the current contents of the commit message which can be validated. If a hook exits nonzero, the commit will be aborted.

Confining hooks to run at certain stages

If pre-commit during push has been installed, then all hooks (by default) will be run during the push stage. Hooks can however be confined to a stage by setting the stages property in your .pre-commit-config.yaml. The stages property is an array and can contain any of commit, push, and commit-msg.

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: v0.9.1
    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 a .pre-commit-config.yaml:

-   repo: git://github.com/path/to/your/hook/repo
    sha: badf00ddeadbeef
    hooks:
    -   id: my-hook-script-id
        args: [--myarg1=1, --myarg1=2]

When you next run pre-commit, your script will be called:

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

If the args property is empty or not defined, your script will be called:

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

Repository Local Hooks

Repository-local hooks are useful when:

  • The scripts are tightly coupled to the repository and it makes sense to distribute the hook scripts with the repository.
  • Hooks require state that is only present in a built artifact of your repository (such as your app's virtualenv for pylint)
  • The official repository for a linter doesn't have the pre-commit metadata.

You can configure repository-local hooks by specifying the repo as the sentinel local.

new in 0.13.0 repository hooks can use any language which supports additional_dependencies or pcre / script / system. This enables you to install things which previously would require a trivial mirror repository.

A local hook must define id, name, language, entry, and files as specified under Creating new hooks

Here's an example configuration with a few local hooks:

-   repo: local
    hooks:
    -   id: pylint
        name: pylint
        entry: python -m pylint.__main__
        language: system
        types: [python]
    -   id: check-x
        name: Check X
        entry: ./bin/check-x.sh
        language: script
        files: \.x$
    -   id: scss-lint
        name: scss-lint
        entry: scss-lint
        language: ruby
        language_version: 2.1.5
        types: [scss]
        additional_dependencies: ['scss_lint:0.52.0']

Filtering files with types

new in 0.15.0

Filtering with types provides several benefits over traditional filtering with files.

  • no error-prone regular expressions
  • files can be matched by their shebang (even when extensionless)
  • symlinks / submodules can be easily ignored

types is specified per hook as an array of tags. The tags are discovered through a set of heuristics by the identify library. identify was chosen as it is a small portable pure python library.

Some of the common tags you'll find from identify:

  • file
  • symlink
  • directory - in the context of pre-commit this will be a submodule
  • executable - whether the file has the executable bit set
  • text - whether the file looks like a text file
  • binary - whether the file looks like a binary file
  • tags by extension / naming convention
  • tags by shebang (#!)

To discovery the type of any file on disk, you can use identify's cli:

$ identify-cli setup.py
["file", "non-executable", "python", "text"]
$ identify-cli some-random-file
["file", "non-executable", "text"]
$ identify-cli --filename-only some-random-file; echo $?
1

If a file extension you use is not supported, please submit a pull request!

types and files are evaluated with AND when filtering. Tags within types are also evaluated using AND.

For example:

    files: ^foo/
    types: [file, python]
will match a file foo/1.py but will not match setup.py.

Files can also be matched by shebang. With types: python, exe starting with #!/usr/bin/env python3 will also be matched.

As with files and exclude, you can also exclude types if necessary using exclude_types.

If you'd like to use types with compatibility for older versions here is a guide to ensuring compatibility.

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: v0.43.2
    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. To check only files which have changed, which may be faster, use something like git diff-tree --no-commit-id --name-only -r $REVISION | xargs pre-commit run --files.

We’re looking to grow the project and get more contributors especially to support more languages/versions. We’d also like to get the .pre-commit-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.