.github/FUNDING.yml

@@ -0,0 +1,2 @@
+github: asottile
+open_collective: pre-commit

.github/ISSUE_TEMPLATE/00_bug.yaml

@@ -1,54 +0,0 @@
-name: bug report
-description: something went wrong
-body:
-  - type: markdown
-    attributes:
-      value: |
-         this is for issues for `pre-commit` (the framework).
-         if you are reporting an issue for [pre-commit.ci] please report it at [pre-commit-ci/issues]
-
-         [pre-commit.ci]: https://pre-commit.ci
-         [pre-commit-ci/issues]: https://github.com/pre-commit-ci/issues
-  - type: input
-    id: search
-    attributes:
-      label: search you tried in the issue tracker
-      placeholder: ...
-    validations:
-      required: true
-  - type: markdown
-    attributes:
-      value: |
-        95% of issues created are duplicates.
-        please try extra hard to find them first.
-        it's very unlikely your problem is unique.
-  - type: textarea
-    id: freeform
-    attributes:
-      label: describe your issue
-      placeholder: 'I was doing ... I ran ... I expected ... I got ...'
-    validations:
-      required: true
-  - type: input
-    id: version
-    attributes:
-      label: pre-commit --version
-      placeholder: pre-commit x.x.x
-    validations:
-      required: true
-  - type: textarea
-    id: configuration
-    attributes:
-      label: .pre-commit-config.yaml
-      description: (auto-rendered as yaml, no need for backticks)
-      placeholder: 'repos: ...'
-      render: yaml
-    validations:
-      required: true
-  - type: textarea
-    id: error-log
-    attributes:
-      label: '~/.cache/pre-commit/pre-commit.log (if present)'
-      placeholder: "### version information\n..."
-    validations:
-      required: false

.github/ISSUE_TEMPLATE/01_feature.yaml

@@ -1,38 +0,0 @@
-name: feature request
-description: something new
-body:
-  - type: markdown
-    attributes:
-      value: |
-         this is for issues for `pre-commit` (the framework).
-         if you are reporting an issue for [pre-commit.ci] please report it at [pre-commit-ci/issues]
-
-         [pre-commit.ci]: https://pre-commit.ci
-         [pre-commit-ci/issues]: https://github.com/pre-commit-ci/issues
-  - type: input
-    id: search
-    attributes:
-      label: search you tried in the issue tracker
-      placeholder: ...
-    validations:
-      required: true
-  - type: markdown
-    attributes:
-      value: |
-        95% of issues created are duplicates.
-        please try extra hard to find them first.
-        it's very unlikely your feature idea is a new one.
-  - type: textarea
-    id: freeform
-    attributes:
-      label: describe your actual problem
-      placeholder: 'I want to do ... I tried ... It does not work because ...'
-    validations:
-      required: true
-  - type: input
-    id: version
-    attributes:
-      label: pre-commit --version
-      placeholder: pre-commit x.x.x
-    validations:
-      required: true

.github/ISSUE_TEMPLATE/config.yml

@@ -1,8 +0,0 @@
-blank_issues_enabled: false
-contact_links:
-- name: documentation
-  url: https://pre-commit.com
-  about: please check the docs first
-- name: pre-commit.ci issues
-  url: https://github.com/pre-commit-ci/issues
-  about: please report issues about pre-commit.ci here

.github/actions/pre-test/action.yml

@@ -1,9 +0,0 @@
-inputs:
-  env:
-    default: ${{ matrix.env }}
-
-runs:
-  using: composite
-  steps:
-  - uses: asottile/workflows/.github/actions/latest-git@v1.4.0
-    if: inputs.env == 'py39' && runner.os == 'Linux'

.github/workflows/languages.yaml

@@ -1,84 +0,0 @@
-name: languages
-
-on:
-  push:
-    branches: [main, test-me-*]
-    tags: '*'
-  pull_request:
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
-  cancel-in-progress: true
-
-jobs:
-  vars:
-    runs-on: ubuntu-latest
-    outputs:
-      languages: ${{ steps.vars.outputs.languages }}
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-    - uses: actions/setup-python@v4
-      with:
-        python-version: '3.10'
-    - name: install deps
-      run: python -mpip install -e . -r requirements-dev.txt
-    - name: vars
-      run: testing/languages ${{ github.event_name == 'push' && '--all' || '' }}
-      id: vars
-  language:
-    needs: [vars]
-    runs-on: ${{ matrix.os }}
-    if: needs.vars.outputs.languages != '[]'
-    strategy:
-      fail-fast: false
-      matrix:
-        include: ${{ fromJSON(needs.vars.outputs.languages) }}
-    steps:
-    - uses: asottile/workflows/.github/actions/fast-checkout@v1.8.1
-    - uses: actions/setup-python@v4
-      with:
-        python-version: '3.10'
-
-    - run: echo "$CONDA\Scripts" >> "$GITHUB_PATH"
-      shell: bash
-      if: matrix.os == 'windows-latest' && matrix.language == 'conda'
-    - run: testing/get-coursier.sh
-      shell: bash
-      if: matrix.language == 'coursier'
-    - run: testing/get-dart.sh
-      shell: bash
-      if: matrix.language == 'dart'
-    - run: |
-        sudo apt-get update
-        sudo apt-get install -y --no-install-recommends \
-          lua5.3 \
-          liblua5.3-dev \
-          luarocks
-      if: matrix.os == 'ubuntu-latest' && matrix.language == 'lua'
-    - run: |
-        echo 'C:\Strawberry\perl\bin' >> "$GITHUB_PATH"
-        echo 'C:\Strawberry\perl\site\bin' >> "$GITHUB_PATH"
-        echo 'C:\Strawberry\c\bin' >> "$GITHUB_PATH"
-      shell: bash
-      if: matrix.os == 'windows-latest' && matrix.language == 'perl'
-    - uses: haskell/actions/setup@v2
-      if: matrix.language == 'haskell'
-    - uses: r-lib/actions/setup-r@v2
-      if: matrix.os == 'ubuntu-latest' && matrix.language == 'r'
-
-    - name: install deps
-      run: python -mpip install -e . -r requirements-dev.txt
-    - name: run tests
-      run: coverage run -m pytest tests/languages/${{ matrix.language }}_test.py
-    - name: check coverage
-      run: coverage report --include pre_commit/languages/${{ matrix.language }}.py,tests/languages/${{ matrix.language }}_test.py
-  collector:
-    needs: [language]
-    if: always()
-    runs-on: ubuntu-latest
-    steps:
-    - name: check for failures
-      if: contains(needs.*.result, 'failure') || contains(needs.*.result, 'cancelled')
-      run: echo job failed && exit 1

.github/workflows/main.yml

@@ -1,23 +0,0 @@
-name: main
-
-on:
-  push:
-    branches: [main, test-me-*]
-    tags: '*'
-  pull_request:
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
-  cancel-in-progress: true
-
-jobs:
-  main-windows:
-    uses: asottile/workflows/.github/workflows/tox.yml@v1.8.1
-    with:
-      env: '["py310"]'
-      os: windows-latest
-  main-linux:
-    uses: asottile/workflows/.github/workflows/tox.yml@v1.8.1
-    with:
-      env: '["py310", "py311", "py312", "py313"]'
-      os: ubuntu-latest

.gitignore

@@ -1,6 +1,9 @@
 *.egg-info
 *.py[co]
 /.coverage
+/.mypy_cache
+/.pytest_cache
 /.tox
 /dist
+/venv*
 .vscode/

.pre-commit-config.yaml

@@ -1,44 +1,55 @@
 repos:
 -   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v6.0.0
+    rev: v4.0.1
     hooks:
     -   id: trailing-whitespace
     -   id: end-of-file-fixer
+    -   id: check-docstring-first
+    -   id: check-json
     -   id: check-yaml
     -   id: debug-statements
-    -   id: double-quote-string-fixer
     -   id: name-tests-test
     -   id: requirements-txt-fixer
--   repo: https://github.com/asottile/setup-cfg-fmt
-    rev: v3.2.0
-    hooks:
-    -   id: setup-cfg-fmt
--   repo: https://github.com/asottile/reorder-python-imports
-    rev: v3.16.0
-    hooks:
-    -   id: reorder-python-imports
-        exclude: ^pre_commit/resources/
-        args: [--py310-plus, --add-import, 'from __future__ import annotations']
--   repo: https://github.com/asottile/add-trailing-comma
-    rev: v4.0.0
-    hooks:
-    -   id: add-trailing-comma
--   repo: https://github.com/asottile/pyupgrade
-    rev: v3.21.2
-    hooks:
-    -   id: pyupgrade
-        args: [--py310-plus]
--   repo: https://github.com/hhatto/autopep8
-    rev: v2.3.2
-    hooks:
-    -   id: autopep8
+    -   id: double-quote-string-fixer
 -   repo: https://github.com/PyCQA/flake8
-    rev: 7.3.0
+    rev: 3.9.2
     hooks:
     -   id: flake8
+        additional_dependencies: [flake8-typing-imports==1.10.0]
+-   repo: https://github.com/pre-commit/mirrors-autopep8
+    rev: v1.5.7
+    hooks:
+    -   id: autopep8
+-   repo: https://github.com/pre-commit/pre-commit
+    rev: v2.15.0
+    hooks:
+    -   id: validate_manifest
+-   repo: https://github.com/asottile/pyupgrade
+    rev: v2.25.0
+    hooks:
+    -   id: pyupgrade
+        args: [--py36-plus]
+-   repo: https://github.com/asottile/reorder_python_imports
+    rev: v2.6.0
+    hooks:
+    -   id: reorder-python-imports
+        args: [--py3-plus]
+-   repo: https://github.com/asottile/add-trailing-comma
+    rev: v2.1.0
+    hooks:
+    -   id: add-trailing-comma
+        args: [--py36-plus]
+-   repo: https://github.com/asottile/setup-cfg-fmt
+    rev: v1.17.0
+    hooks:
+    -   id: setup-cfg-fmt
 -   repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v1.19.1
+    rev: v0.910
     hooks:
     -   id: mypy
-        additional_dependencies: [types-pyyaml]
+        additional_dependencies: [types-all]
         exclude: ^testing/resources/
+-   repo: meta
+    hooks:
+    -   id: check-hooks-apply
+    -   id: check-useless-excludes

.pre-commit-hooks.yaml

@@ -1,6 +1,6 @@
 -   id: validate_manifest
     name: validate pre-commit manifest
     description: This validator validates a pre-commit hooks manifest file
-    entry: pre-commit validate-manifest
+    entry: pre-commit-validate-manifest
     language: python
-    files: ^\.pre-commit-hooks\.yaml$
+    files: ^(\.pre-commit-hooks\.yaml|hooks\.yaml)$

CHANGELOG.md

@@ -1,579 +1,3 @@
-4.5.1 - 2025-12-16
-==================
-
-### Fixes
-- Fix `language: python` with `repo: local` without `additional_dependencies`.
-    - #3597 PR by @asottile.
-
-4.5.0 - 2025-11-22
-==================
-
-### Features
-- Add `pre-commit hazmat`.
-    - #3585 PR by @asottile.
-
-4.4.0 - 2025-11-08
-==================
-
-### Features
-- Add `--fail-fast` option to `pre-commit run`.
-    - #3528 PR by @JulianMaurin.
-- Upgrade `ruby-build` / `rbenv`.
-    - #3566 PR by @asottile.
-    - #3565 issue by @MRigal.
-- Add `language: unsupported` / `language: unsupported_script` as aliases
-  for `language: system` / `language: script` (which will eventually be
-  deprecated).
-    - #3577 PR by @asottile.
-- Add support docker-in-docker detection for cgroups v2.
-    - #3535 PR by @br-rhrbacek.
-    - #3360 issue by @JasonAlt.
-
-### Fixes
-- Handle when docker gives `SecurityOptions: null`.
-    - #3537 PR by @asottile.
-    - #3514 issue by @jenstroeger.
-- Fix error context for invalid `stages` in `.pre-commit-config.yaml`.
-    - #3576 PR by @asottile.
-
-4.3.0 - 2025-08-09
-==================
-
-### Features
-- `language: docker` / `language: docker_image`: detect rootless docker.
-    - #3446 PR by @matthewhughes934.
-    - #1243 issue by @dkolepp.
-- `language: julia`: avoid `startup.jl` when executing hooks.
-    - #3496 PR by @ericphanson.
-- `language: dart`: support latest dart versions which require a higher sdk
-  lower bound.
-    - #3507 PR by @bc-lee.
-
-4.2.0 - 2025-03-18
-==================
-
-### Features
-- For `language: python` first attempt a versioned python executable for
-  the default language version before consulting a potentially unversioned
-  `sys.executable`.
-    - #3430 PR by @asottile.
-
-### Fixes
-- Handle error during conflict detection when a file is named "HEAD"
-    - #3425 PR by @tusharsadhwani.
-
-4.1.0 - 2025-01-20
-==================
-
-### Features
-- Add `language: julia`.
-    - #3348 PR by @fredrikekre.
-    - #2689 issue @jmuchovej.
-
-### Fixes
-- Disable automatic toolchain switching for `language: golang`.
-    - #3304 PR by @AleksaC.
-    - #3300 issue by @AleksaC.
-    - #3149 issue by @nijel.
-- Fix `language: r` installation when initiated by RStudio.
-    - #3389 PR by @lorenzwalthert.
-    - #3385 issue by @lorenzwalthert.
-
-
-4.0.1 - 2024-10-08
-==================
-
-### Fixes
-- Fix `pre-commit migrate-config` for unquoted deprecated stages names with
-  purelib `pyyaml`.
-    - #3324 PR by @asottile.
-    - pre-commit-ci/issues#234 issue by @lorenzwalthert.
-
-4.0.0 - 2024-10-05
-==================
-
-### Features
-- Improve `pre-commit migrate-config` to handle more yaml formats.
-    - #3301 PR by @asottile.
-- Handle `stages` deprecation in `pre-commit migrate-config`.
-    - #3302 PR by @asottile.
-    - #2732 issue by @asottile.
-- Upgrade `ruby-build`.
-    - #3199 PR by @ThisGuyCodes.
-- Add "sensible regex" warnings to `repo: meta`.
-    - #3311 PR by @asottile.
-- Add warnings for deprecated `stages` (`commit` -> `pre-commit`, `push` ->
-  `pre-push`, `merge-commit` -> `pre-merge-commit`).
-    - #3312 PR by @asottile.
-    - #3313 PR by @asottile.
-    - #3315 PR by @asottile.
-    - #2732 issue by @asottile.
-
-### Updating
-- `language: python_venv` has been removed -- use `language: python` instead.
-    - #3320 PR by @asottile.
-    - #2734 issue by @asottile.
-
-3.8.0 - 2024-07-28
-==================
-
-### Features
-- Implement health checks for `language: r` so environments are recreated if
-  the system version of R changes.
-    - #3206 issue by @lorenzwalthert.
-    - #3265 PR by @lorenzwalthert.
-
-3.7.1 - 2024-05-10
-==================
-
-### Fixes
-- Fix `language: rust` default language version check when `rust-toolchain.toml`
-  is present.
-    - issue by @gaborbernat.
-    - #3201 PR by @asottile.
-
-3.7.0 - 2024-03-24
-==================
-
-### Features
-- Use a tty for `docker` and `docker_image` hooks when `--color` is specified.
-    - #3122 PR by @glehmann.
-
-### Fixes
-- Fix `fail_fast` for individual hooks stopping when previous hooks had failed.
-    - #3167 issue by @tp832944.
-    - #3168 PR by @asottile.
-
-### Updating
-- The per-hook behaviour of `fail_fast` was fixed.  If you want the pre-3.7.0
-  behaviour, add `fail_fast: true` to all hooks before the last `fail_fast`
-  hook.
-
-3.6.2 - 2024-02-18
-==================
-
-### Fixes
-- Fix building golang hooks during `git commit --all`.
-    - #3130 PR by @asottile.
-    - #2722 issue by @pestanko and @matthewhughes934.
-
-3.6.1 - 2024-02-10
-==================
-
-### Fixes
-- Remove `PYTHONEXECUTABLE` from environment when running.
-    - #3110 PR by @untitaker.
-- Handle staged-files-only with only a crlf diff.
-    - #3126 PR by @asottile.
-    - issue by @tyyrok.
-
-3.6.0 - 2023-12-09
-==================
-
-### Features
-- Check `minimum_pre_commit_version` first when parsing configs.
-    - #3092 PR by @asottile.
-
-### Fixes
-- Fix deprecation warnings for `importlib.resources`.
-    - #3043 PR by @asottile.
-- Fix deprecation warnings for rmtree.
-    - #3079 PR by @edgarrmondragon.
-
-### Updating
-- Drop support for python<3.9.
-    - #3042 PR by @asottile.
-    - #3093 PR by @asottile.
-
-3.5.0 - 2023-10-13
-==================
-
-### Features
-- Improve performance of `check-hooks-apply` and `check-useless-excludes`.
-    - #2998 PR by @mxr.
-    - #2935 issue by @mxr.
-
-### Fixes
-- Use `time.monotonic()` for more accurate hook timing.
-    - #3024 PR by @adamchainz.
-
-### Updating
-- Require npm 6.x+ for `language: node` hooks.
-    - #2996 PR by @RoelAdriaans.
-    - #1983 issue by @henryiii.
-
-3.4.0 - 2023-09-02
-==================
-
-### Features
-- Add `language: haskell`.
-    - #2932 by @alunduil.
-- Improve cpu count detection when run under cgroups.
-    - #2979 PR by @jdb8.
-    - #2978 issue by @jdb8.
-
-### Fixes
-- Handle negative exit codes from hooks receiving posix signals.
-    - #2971 PR by @chriskuehl.
-    - #2970 issue by @chriskuehl.
-
-3.3.3 - 2023-06-13
-==================
-
-### Fixes
-- Work around OS packagers setting `--install-dir` / `--bin-dir` in gem settings.
-    - #2905 PR by @jaysoffian.
-    - #2799 issue by @lmilbaum.
-
-3.3.2 - 2023-05-17
-==================
-
-### Fixes
-- Work around `r` on windows sometimes double-un-quoting arguments.
-    - #2885 PR by @lorenzwalthert.
-    - #2870 issue by @lorenzwalthert.
-
-3.3.1 - 2023-05-02
-==================
-
-### Fixes
-- Work around `git` partial clone bug for `autoupdate` on windows.
-    - #2866 PR by @asottile.
-    - #2865 issue by @adehad.
-
-3.3.0 - 2023-05-01
-==================
-
-### Features
-- Upgrade ruby-build.
-    - #2846 PR by @jalessio.
-- Use blobless clone for faster autoupdate.
-    - #2859 PR by @asottile.
-- Add `-j` / `--jobs` argument to `autoupdate` for parallel execution.
-    - #2863 PR by @asottile.
-    - issue by @gaborbernat.
-
-3.2.2 - 2023-04-03
-==================
-
-### Fixes
-- Fix support for swift >= 5.8.
-    - #2836 PR by @edelabar.
-    - #2835 issue by @kgrobelny-intive.
-
-3.2.1 - 2023-03-25
-==================
-
-### Fixes
-- Fix `language_version` for `language: rust` without global `rustup`.
-    - #2823 issue by @daschuer.
-    - #2827 PR by @asottile.
-
-3.2.0 - 2023-03-17
-==================
-
-### Features
-- Allow `pre-commit`, `pre-push`, and `pre-merge-commit` as `stages`.
-    - #2732 issue by @asottile.
-    - #2808 PR by @asottile.
-- Add `pre-rebase` hook support.
-    - #2582 issue by @BrutalSimplicity.
-    - #2725 PR by @mgaligniana.
-
-### Fixes
-- Remove bulky cargo cache from `language: rust` installs.
-    - #2820 PR by @asottile.
-
-3.1.1 - 2023-02-27
-==================
-
-### Fixes
-- Fix `rust` with `language_version` and a non-writable host `RUSTUP_HOME`.
-    - pre-commit-ci/issues#173 by @Swiftb0y.
-    - #2788 by @asottile.
-
-3.1.0 - 2023-02-22
-==================
-
-### Fixes
-- Fix `dotnet` for `.sln`-based hooks for dotnet>=7.0.200.
-    - #2763 PR by @m-rsha.
-- Prevent stashing when `diff` fails to execute.
-    - #2774 PR by @asottile.
-    - #2773 issue by @strubbly.
-- Dependencies are no longer sorted in repository key.
-    - #2776 PR by @asottile.
-
-### Updating
-- Deprecate `language: python_venv`.  Use `language: python` instead.
-    - #2746 PR by @asottile.
-    - #2734 issue by @asottile.
-
-
-3.0.4 - 2023-02-03
-==================
-
-### Fixes
-- Fix hook diff detection for files affected by `--textconv`.
-    - #2743 PR by @adamchainz.
-    - #2743 issue by @adamchainz.
-
-3.0.3 - 2023-02-01
-==================
-
-### Fixes
-- Revert "Prevent local `Gemfile` from interfering with hook execution.".
-    - #2739 issue by @Roguelazer.
-    - #2740 PR by @asottile.
-
-3.0.2 - 2023-01-29
-==================
-
-### Fixes
-- Prevent local `Gemfile` from interfering with hook execution.
-    - #2727 PR by @asottile.
-- Fix `language: r`, `repo: local` hooks
-    - pre-commit-ci/issues#107 by @lorenzwalthert.
-    - #2728 PR by @asottile.
-
-3.0.1 - 2023-01-26
-==================
-
-### Fixes
-- Ensure coursier hooks are available offline after install.
-    - #2723 PR by @asottile.
-
-3.0.0 - 2023-01-23
-==================
-
-### Features
-- Make `language: golang` bootstrap `go` if not present.
-    - #2651 PR by @taoufik07.
-    - #2649 issue by @taoufik07.
-- `language: coursier` now supports `additional_dependencies` and `repo: local`
-    - #2702 PR by @asottile.
-- Upgrade `ruby-build` to `20221225`.
-    - #2718 PR by @jalessio.
-
-### Fixes
-- Improve error message for invalid yaml for `pre-commit autoupdate`.
-    - #2686 PR by @asottile.
-    - #2685 issue by @CarstenGrohmann.
-- `repo: local` no longer provisions an empty `git` repo.
-    - #2699 PR by @asottile.
-
-### Updating
-- Drop support for python<3.8
-    - #2655 PR by @asottile.
-- Drop support for top-level list, use `pre-commit migrate-config` to update.
-    - #2656 PR by @asottile.
-- Drop support for `sha` to specify revision, use `pre-commit migrate-config`
-  to update.
-    - #2657 PR by @asottile.
-- Remove `pre-commit-validate-config` and `pre-commit-validate-manifest`, use
-  `pre-commit validate-config` and `pre-commit validate-manifest` instead.
-    - #2658 PR by @asottile.
-- `language: golang` hooks must use `go.mod` to specify dependencies
-    - #2672 PR by @taoufik07.
-
-
-2.21.0 - 2022-12-25
-===================
-
-### Features
-- Require new-enough virtualenv to prevent 3.10 breakage
-    - #2467 PR by @asottile.
-- Respect aliases with `SKIP` for environment install.
-    - #2480 PR by @kmARC.
-    - #2478 issue by @kmARC.
-- Allow `pre-commit run --files` against unmerged paths.
-    - #2484 PR by @asottile.
-- Also apply regex warnings to `repo: local` hooks.
-    - #2524 PR by @chrisRedwine.
-    - #2521 issue by @asottile.
-- `rust` is now a "first class" language -- supporting `language_version` and
-  installation when not present.
-    - #2534 PR by @Holzhaus.
-- `r` now uses more-reliable binary installation.
-    - #2460 PR by @lorenzwalthert.
-- `GIT_ALLOW_PROTOCOL` is now passed through for git operations.
-    - #2555 PR by @asottile.
-- `GIT_ASKPASS` is now passed through for git operations.
-    - #2564 PR by @mattp-.
-- Remove `toml` dependency by using `cargo add` directly.
-    - #2568 PR by @m-rsha.
-- Support `dotnet` hooks which have dotted prefixes.
-    - #2641 PR by @rkm.
-    - #2629 issue by @rkm.
-
-### Fixes
-- Properly adjust `--commit-msg-filename` if run from a sub directory.
-    - #2459 PR by @asottile.
-- Simplify `--intent-to-add` detection by using `git diff`.
-    - #2580 PR by @m-rsha.
-- Fix `R.exe` selection on windows.
-    - #2605 PR by @lorenzwalthert.
-    - #2599 issue by @SInginc.
-- Skip default `nuget` source when installing `dotnet` packages.
-    - #2642 PR by @rkm.
-
-2.20.0 - 2022-07-10
-===================
-
-### Features
-- Expose `source` and `object-name` (positional args) of `prepare-commit-msg`
-  hook as `PRE_COMMIT_COMIT_MSG_SOURCE` and `PRE_COMMIT_COMMIT_OBJECT_NAME`.
-    - #2407 PR by @M-Whitaker.
-    - #2406 issue by @M-Whitaker.
-
-### Fixes
-- Fix `language: ruby` installs when `--user-install` is set in gemrc.
-    - #2394 PR by @narpfel.
-    - #2393 issue by @narpfel.
-- Adjust pty setup for solaris.
-    - #2390 PR by @gaige.
-    - #2389 issue by @gaige.
-- Remove unused `--config` option from `gc`, `sample-config`,
-  `validate-config`, `validate-manifest` sub-commands.
-    - #2429 PR by @asottile.
-
-2.19.0 - 2022-05-05
-===================
-
-### Features
-- Allow multiple outputs from `language: dotnet` hooks.
-    - #2332 PR by @WallucePinkham.
-- Add more information to `healthy()` failure.
-    - #2348 PR by @asottile.
-- Upgrade ruby-build.
-    - #2342 PR by @jalessio.
-- Add `pre-commit validate-config` / `pre-commit validate-manifest` and
-  deprecate `pre-commit-validate-config` and `pre-commit-validate-manifest`.
-    - #2362 PR by @asottile.
-
-### Fixes
-- Fix `pre-push` when pushed ref contains spaces.
-    - #2345 PR by @wwade.
-    - #2344 issue by @wwade.
-
-### Updating
-- Change `pre-commit-validate-config` / `pre-commit-validate-manifest` to
-  `pre-commit validate-config` / `pre-commit validate-manifest`.
-    - #2362 PR by @asottile.
-
-2.18.1 - 2022-04-02
-===================
-
-### Fixes
-- Fix regression for `repo: local` hooks running `python<3.7`
-    - #2324 PR by @asottile.
-
-2.18.0 - 2022-04-02
-===================
-
-### Features
-- Keep `GIT_HTTP_PROXY_AUTHMETHOD` in git environ.
-    - #2272 PR by @VincentBerthier.
-    - #2271 issue by @VincentBerthier.
-- Support both `cs` and `coursier` executables for coursier hooks.
-    - #2293 PR by @Holzhaus.
-- Include more information in errors for `language_version` /
-  `additional_dependencies` for languages which do not support them.
-    -  #2315 PR by @asottile.
-- Have autoupdate preferentially pick tags which look like versions when
-  there are multiple equivalent tags.
-    - #2312 PR by @mblayman.
-    - #2311 issue by @mblayman.
-- Upgrade `ruby-build`.
-    - #2319 PR by @jalessio.
-- Add top level `default_install_hook_types` which will be installed when
-  `--hook-types` is not specified in `pre-commit install`.
-    - #2322 PR by @asottile.
-
-### Fixes
-- Fix typo in help message for `--from-ref` and `--to-ref`.
-    - #2266 PR by @leetrout.
-- Prioritize binary builds for R dependencies.
-    - #2277 PR by @lorenzwalthert.
-- Fix handling of git worktrees.
-    - #2252 PR by @daschuer.
-- Fix handling of `$R_HOME` for R hooks.
-    - #2301 PR by @jeff-m-sullivan.
-    - #2300 issue by @jeff-m-sullivan.
-- Fix a rare race condition in change stashing.
-    - #2323 PR by @asottile.
-    - #2287 issue by @ian-h-chamberlain.
-
-### Updating
-- Remove python3.6 support.  Note that pre-commit still supports running hooks
-  written in older versions, but pre-commit itself requires python 3.7+.
-    - #2215 PR by @asottile.
-- pre-commit has migrated from the `master` branch to `main`.
-    - #2302 PR by @asottile.
-
-2.17.0 - 2022-01-18
-===================
-
-### Features
-- add warnings for regexes containing `[\\/]`.
-    - #2151 issue by @sanjioh.
-    - #2154 PR by @kuviokelluja.
-- upgrade supported ruby versions.
-    - #2205 PR by @jalessio.
-- allow `language: conda` to use `mamba` or `micromamba` via
-  `PRE_COMMIT_USE_MAMBA=1` or `PRE_COMMIT_USE_MICROMAMBA=1` respectively.
-    - #2204 issue by @janjagusch.
-    - #2207 PR by @xhochy.
-- display `git --version` in error report.
-    - #2210 PR by @asottile.
-- add `language: lua` as a supported language.
-    - #2158 PR by @mblayman.
-
-### Fixes
-- temporarily add `setuptools` to the zipapp.
-    - #2122 issue by @andreoliwa.
-    - a737d5f commit by @asottile.
-- use `go install` instead of `go get` for go 1.18+ support.
-    - #2161 PR by @schmir.
-- fix `language: r` with a local renv and `RENV_PROJECT` set.
-    - #2170 PR by @lorenzwalthert.
-- forbid overriding `entry` in `language: meta` hooks which breaks them.
-    - #2180 issue by @DanKaplanSES.
-    - #2181 PR by @asottile.
-- always use `#!/bin/sh` on windows for hook script.
-    - #2182 issue by @hushigome-visco.
-    - #2187 PR by @asottile.
-
-2.16.0 - 2021-11-30
-===================
-
-### Features
-- add warning for regexes containing `[\/]` or `[/\\]`.
-    - #2053 PR by @radek-sprta.
-    - #2043 issue by @asottile.
-- move hook template back to `bash` resolving shebang-portability issues.
-    - #2065 PR by @asottile.
-- add support for `fail_fast` at the individual hook level.
-    - #2097 PR by @colens3.
-    - #1143 issue by @potiuk.
-- allow passthrough of `GIT_CONFIG_KEY_*`, `GIT_CONFIG_VALUE_*`, and
-  `GIT_CONFIG_COUNT`.
-    - #2136 PR by @emzeat.
-
-### Fixes
-- fix pre-commit autoupdate for `core.useBuiltinFSMonitor=true` on windows.
-    - #2047 PR by @asottile.
-    - #2046 issue by @lcnittl.
-- fix temporary file stashing with for `submodule.recurse=1`.
-    - #2071 PR by @asottile.
-    - #2063 issue by @a666.
-- ban broken importlib-resources versions.
-    - #2098 PR by @asottile.
-- replace `exit(...)` with `raise SystemExit(...)` for portability.
-    - #2103 PR by @asottile.
-    - #2104 PR by @asottile.
-
-
 2.15.0 - 2021-09-02
 ===================
 

CONTRIBUTING.md

@@ -5,6 +5,7 @@
 - The complete test suite depends on having at least the following installed
   (possibly not a complete list)
   - git (Version 2.24.0 or above is required to run pre-merge-commit tests)
+  - python2 (Required by a test which checks different python versions)
   - python3 (Required by a test which checks different python versions)
   - tox (or virtualenv)
   - ruby + gem
@@ -64,14 +65,14 @@ to implement.  The current implemented languages are at varying levels:
 - 0th class - pre-commit does not require any dependencies for these languages
   as they're not actually languages (current examples: fail, pygrep)
 - 1st class - pre-commit will bootstrap a full interpreter requiring nothing to
-  be installed globally (current examples: go, node, ruby, rust)
+  be installed globally (current examples: node, ruby)
 - 2nd class - pre-commit requires the user to install the language globally but
-  will install tools in an isolated fashion (current examples: python, swift,
-  docker).
+  will install tools in an isolated fashion (current examples: python, go, rust,
+  swift, docker).
 - 3rd class - pre-commit requires the user to install both the tool and the
   language globally (current examples: script, system)
 
-"second class" is usually the easiest to implement first and is perfectly
+"third class" is usually the easiest to implement first and is perfectly
 acceptable.
 
 Ideally the language works on the supported platforms for pre-commit (linux,
@@ -92,7 +93,7 @@ language, for example:
 
 here are the apis that should be implemented for a language
 
-Note that these are also documented in [`pre_commit/lang_base.py`](https://github.com/pre-commit/pre-commit/blob/main/pre_commit/lang_base.py)
+Note that these are also documented in [`pre_commit/languages/all.py`](https://github.com/pre-commit/pre-commit/blob/master/pre_commit/languages/all.py)
 
 #### `ENVIRONMENT_DIR`
 
@@ -111,21 +112,20 @@ one cannot be determined, return `'default'`.
 You generally don't need to implement this on a first pass and can just use:
 
 ```python
-get_default_version = lang_base.basic_default_version
+get_default_version = helpers.basic_default_version
 ```
 
 `python` is currently the only language which implements this api
 
-#### `health_check`
+#### `healthy`
 
 This is used to check whether the installed environment is considered healthy.
-This function should return a detailed message if unhealthy or `None` if
-healthy.
+This function should return `True` or `False`.
 
 You generally don't need to implement this on a first pass and can just use:
 
 ```python
-health_check = lang_base.basic_health_check
+healthy = helpers.basic_healthy
 ```
 
 `python` is currently the only language which implements this api, for python
@@ -137,7 +137,7 @@ this is the trickiest one to implement and where all the smart parts happen.
 
 this api should do the following things
 
-- (0th / 3rd class): `install_environment = lang_base.no_install`
+- (0th / 3rd class): `install_environment = helpers.no_install`
 - (1st class): install a language runtime into the hook's directory
 - (2nd class): install the package at `.` into the `ENVIRONMENT_DIR`
 - (2nd class, optional): install packages listed in `additional_dependencies`

README.md

@@ -1,5 +1,6 @@
-[![build status](https://github.com/pre-commit/pre-commit/actions/workflows/main.yml/badge.svg)](https://github.com/pre-commit/pre-commit/actions/workflows/main.yml)
-[![pre-commit.ci status](https://results.pre-commit.ci/badge/github/pre-commit/pre-commit/main.svg)](https://results.pre-commit.ci/latest/github/pre-commit/pre-commit/main)
+[![Build Status](https://dev.azure.com/asottile/asottile/_apis/build/status/pre-commit.pre-commit?branchName=master)](https://dev.azure.com/asottile/asottile/_build/latest?definitionId=21&branchName=master)
+[![Azure DevOps coverage](https://img.shields.io/azure-devops/coverage/asottile/asottile/21/master.svg)](https://dev.azure.com/asottile/asottile/_build/latest?definitionId=21&branchName=master)
+[![pre-commit.ci status](https://results.pre-commit.ci/badge/github/pre-commit/pre-commit/master.svg)](https://results.pre-commit.ci/latest/github/pre-commit/pre-commit/master)
 
 ## pre-commit
 

azure-pipelines.yml

@@ -0,0 +1,62 @@
+trigger:
+  branches:
+    include: [master, test-me-*]
+  tags:
+    include: ['*']
+
+resources:
+  repositories:
+    - repository: asottile
+      type: github
+      endpoint: github
+      name: asottile/azure-pipeline-templates
+      ref: refs/tags/v2.1.0
+
+jobs:
+- template: job--python-tox.yml@asottile
+  parameters:
+    toxenvs: [py37]
+    os: windows
+    pre_test:
+    - task: UseRubyVersion@0
+    - powershell: Write-Host "##vso[task.prependpath]$env:CONDA\Scripts"
+      displayName: Add conda to PATH
+    - powershell: |
+        Write-Host "##vso[task.prependpath]C:\Strawberry\perl\bin"
+        Write-Host "##vso[task.prependpath]C:\Strawberry\perl\site\bin"
+        Write-Host "##vso[task.prependpath]C:\Strawberry\c\bin"
+      displayName: Add strawberry perl to PATH
+    - bash: testing/get-dart.sh
+      displayName: install dart
+    - powershell: testing/get-r.ps1
+      displayName: install R
+- template: job--python-tox.yml@asottile
+  parameters:
+    toxenvs: [py37]
+    os: linux
+    name_postfix: _latest_git
+    pre_test:
+    - task: UseRubyVersion@0
+    - template: step--git-install.yml
+    - bash: testing/get-coursier.sh
+      displayName: install coursier
+    - bash: testing/get-dart.sh
+      displayName: install dart
+    - bash: testing/get-swift.sh
+      displayName: install swift
+    - bash: testing/get-r.sh
+      displayName: install R
+- template: job--python-tox.yml@asottile
+  parameters:
+    toxenvs: [pypy3, py36, py37, py38, py39]
+    os: linux
+    pre_test:
+    - task: UseRubyVersion@0
+    - bash: testing/get-coursier.sh
+      displayName: install coursier
+    - bash: testing/get-dart.sh
+      displayName: install dart
+    - bash: testing/get-swift.sh
+      displayName: install swift
+    - bash: testing/get-r.sh
+      displayName: install R

pre_commit/__main__.py

@@ -1,7 +1,5 @@
-from __future__ import annotations
-
 from pre_commit.main import main
 
 
 if __name__ == '__main__':
-    raise SystemExit(main())
+    exit(main())

pre_commit/all_languages.py

@@ -1,50 +0,0 @@
-from __future__ import annotations
-
-from pre_commit.lang_base import Language
-from pre_commit.languages import conda
-from pre_commit.languages import coursier
-from pre_commit.languages import dart
-from pre_commit.languages import docker
-from pre_commit.languages import docker_image
-from pre_commit.languages import dotnet
-from pre_commit.languages import fail
-from pre_commit.languages import golang
-from pre_commit.languages import haskell
-from pre_commit.languages import julia
-from pre_commit.languages import lua
-from pre_commit.languages import node
-from pre_commit.languages import perl
-from pre_commit.languages import pygrep
-from pre_commit.languages import python
-from pre_commit.languages import r
-from pre_commit.languages import ruby
-from pre_commit.languages import rust
-from pre_commit.languages import swift
-from pre_commit.languages import unsupported
-from pre_commit.languages import unsupported_script
-
-
-languages: dict[str, Language] = {
-    'conda': conda,
-    'coursier': coursier,
-    'dart': dart,
-    'docker': docker,
-    'docker_image': docker_image,
-    'dotnet': dotnet,
-    'fail': fail,
-    'golang': golang,
-    'haskell': haskell,
-    'julia': julia,
-    'lua': lua,
-    'node': node,
-    'perl': perl,
-    'pygrep': pygrep,
-    'python': python,
-    'r': r,
-    'ruby': ruby,
-    'rust': rust,
-    'swift': swift,
-    'unsupported': unsupported,
-    'unsupported_script': unsupported_script,
-}
-language_names = sorted(languages)

pre_commit/clientlib.py

@@ -1,43 +1,29 @@
-from __future__ import annotations
-
+import argparse
 import functools
 import logging
-import os.path
 import re
 import shlex
 import sys
-from collections.abc import Callable
-from collections.abc import Sequence
 from typing import Any
-from typing import NamedTuple
+from typing import Dict
+from typing import Optional
+from typing import Sequence
 
 import cfgv
 from identify.identify import ALL_TAGS
 
 import pre_commit.constants as C
-from pre_commit.all_languages import language_names
+from pre_commit.color import add_color_option
 from pre_commit.errors import FatalError
-from pre_commit.yaml import yaml_load
+from pre_commit.languages.all import all_languages
+from pre_commit.logging_handler import logging_handler
+from pre_commit.util import parse_version
+from pre_commit.util import yaml_load
 
 logger = logging.getLogger('pre_commit')
 
 check_string_regex = cfgv.check_and(cfgv.check_string, cfgv.check_regex)
 
-HOOK_TYPES = (
-    'commit-msg',
-    'post-checkout',
-    'post-commit',
-    'post-merge',
-    'post-rewrite',
-    'pre-commit',
-    'pre-merge-commit',
-    'pre-push',
-    'pre-rebase',
-    'prepare-commit-msg',
-)
-# `manual` is not invoked by any installed git hook.  See #719
-STAGES = (*HOOK_TYPES, 'manual')
-
 
 def check_type_tag(tag: str) -> None:
     if tag not in ALL_TAGS:
@@ -47,11 +33,6 @@ def check_type_tag(tag: str) -> None:
         )
 
 
-def parse_version(s: str) -> tuple[int, ...]:
-    """poor man's version comparison"""
-    return tuple(int(p) for p in s.split('.'))
-
-
 def check_min_version(version: str) -> None:
     if parse_version(version) > parse_version(C.VERSION):
         raise cfgv.ValidationError(
@@ -61,186 +42,21 @@ def check_min_version(version: str) -> None:
         )
 
 
-_STAGES = {
-    'commit': 'pre-commit',
-    'merge-commit': 'pre-merge-commit',
-    'push': 'pre-push',
-}
-
-
-def transform_stage(stage: str) -> str:
-    return _STAGES.get(stage, stage)
-
-
-MINIMAL_MANIFEST_SCHEMA = cfgv.Array(
-    cfgv.Map(
-        'Hook', 'id',
-        cfgv.Required('id', cfgv.check_string),
-        cfgv.Optional('stages', cfgv.check_array(cfgv.check_string), []),
-    ),
-)
-
-
-def warn_for_stages_on_repo_init(repo: str, directory: str) -> None:
-    try:
-        manifest = cfgv.load_from_filename(
-            os.path.join(directory, C.MANIFEST_FILE),
-            schema=MINIMAL_MANIFEST_SCHEMA,
-            load_strategy=yaml_load,
-            exc_tp=InvalidManifestError,
-        )
-    except InvalidManifestError:
-        return  # they'll get a better error message when it actually loads!
-
-    legacy_stages = {}  # sorted set
-    for hook in manifest:
-        for stage in hook.get('stages', ()):
-            if stage in _STAGES:
-                legacy_stages[stage] = True
-
-    if legacy_stages:
-        logger.warning(
-            f'repo `{repo}` uses deprecated stage names '
-            f'({", ".join(legacy_stages)}) which will be removed in a '
-            f'future version.  '
-            f'Hint: often `pre-commit autoupdate --repo {shlex.quote(repo)}` '
-            f'will fix this.  '
-            f'if it does not -- consider reporting an issue to that repo.',
-        )
-
-
-class StagesMigrationNoDefault(NamedTuple):
-    key: str
-    default: Sequence[str]
-
-    def check(self, dct: dict[str, Any]) -> None:
-        if self.key not in dct:
-            return
-
-        with cfgv.validate_context(f'At key: {self.key}'):
-            val = dct[self.key]
-            cfgv.check_array(cfgv.check_any)(val)
-
-            val = [transform_stage(v) for v in val]
-            cfgv.check_array(cfgv.check_one_of(STAGES))(val)
-
-    def apply_default(self, dct: dict[str, Any]) -> None:
-        if self.key not in dct:
-            return
-        dct[self.key] = [transform_stage(v) for v in dct[self.key]]
-
-    def remove_default(self, dct: dict[str, Any]) -> None:
-        raise NotImplementedError
-
-
-class StagesMigration(StagesMigrationNoDefault):
-    def apply_default(self, dct: dict[str, Any]) -> None:
-        dct.setdefault(self.key, self.default)
-        super().apply_default(dct)
-
-
-class DeprecatedStagesWarning(NamedTuple):
-    key: str
-
-    def check(self, dct: dict[str, Any]) -> None:
-        if self.key not in dct:
-            return
-
-        val = dct[self.key]
-        cfgv.check_array(cfgv.check_any)(val)
-
-        legacy_stages = [stage for stage in val if stage in _STAGES]
-        if legacy_stages:
-            logger.warning(
-                f'hook id `{dct["id"]}` uses deprecated stage names '
-                f'({", ".join(legacy_stages)}) which will be removed in a '
-                f'future version.  '
-                f'run: `pre-commit migrate-config` to automatically fix this.',
-            )
-
-    def apply_default(self, dct: dict[str, Any]) -> None:
-        pass
-
-    def remove_default(self, dct: dict[str, Any]) -> None:
-        raise NotImplementedError
-
-
-class DeprecatedDefaultStagesWarning(NamedTuple):
-    key: str
-
-    def check(self, dct: dict[str, Any]) -> None:
-        if self.key not in dct:
-            return
-
-        val = dct[self.key]
-        cfgv.check_array(cfgv.check_any)(val)
-
-        legacy_stages = [stage for stage in val if stage in _STAGES]
-        if legacy_stages:
-            logger.warning(
-                f'top-level `default_stages` uses deprecated stage names '
-                f'({", ".join(legacy_stages)}) which will be removed in a '
-                f'future version.  '
-                f'run: `pre-commit migrate-config` to automatically fix this.',
-            )
-
-    def apply_default(self, dct: dict[str, Any]) -> None:
-        pass
-
-    def remove_default(self, dct: dict[str, Any]) -> None:
-        raise NotImplementedError
-
-
-def _translate_language(name: str) -> str:
-    return {
-        'system': 'unsupported',
-        'script': 'unsupported_script',
-    }.get(name, name)
-
-
-class LanguageMigration(NamedTuple):  # remove
-    key: str
-    check_fn: Callable[[object], None]
-
-    def check(self, dct: dict[str, Any]) -> None:
-        if self.key not in dct:
-            return
-
-        with cfgv.validate_context(f'At key: {self.key}'):
-            self.check_fn(_translate_language(dct[self.key]))
-
-    def apply_default(self, dct: dict[str, Any]) -> None:
-        if self.key not in dct:
-            return
-
-        dct[self.key] = _translate_language(dct[self.key])
-
-    def remove_default(self, dct: dict[str, Any]) -> None:
-        raise NotImplementedError
-
-
-class LanguageMigrationRequired(LanguageMigration):  # replace with Required
-    def check(self, dct: dict[str, Any]) -> None:
-        if self.key not in dct:
-            raise cfgv.ValidationError(f'Missing required key: {self.key}')
-
-        super().check(dct)
+def _make_argparser(filenames_help: str) -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser()
+    parser.add_argument('filenames', nargs='*', help=filenames_help)
+    parser.add_argument('-V', '--version', action='version', version=C.VERSION)
+    add_color_option(parser)
+    return parser
 
 
 MANIFEST_HOOK_DICT = cfgv.Map(
     'Hook', 'id',
 
-    # check first in case it uses some newer, incompatible feature
-    cfgv.Optional(
-        'minimum_pre_commit_version',
-        cfgv.check_and(cfgv.check_string, check_min_version),
-        '0',
-    ),
-
     cfgv.Required('id', cfgv.check_string),
     cfgv.Required('name', cfgv.check_string),
     cfgv.Required('entry', cfgv.check_string),
-    LanguageMigrationRequired('language', cfgv.check_one_of(language_names)),
+    cfgv.Required('language', cfgv.check_one_of(all_languages)),
     cfgv.Optional('alias', cfgv.check_string, ''),
 
     cfgv.Optional('files', check_string_regex, ''),
@@ -254,13 +70,13 @@ MANIFEST_HOOK_DICT = cfgv.Map(
     ),
     cfgv.Optional('args', cfgv.check_array(cfgv.check_string), []),
     cfgv.Optional('always_run', cfgv.check_bool, False),
-    cfgv.Optional('fail_fast', cfgv.check_bool, False),
     cfgv.Optional('pass_filenames', cfgv.check_bool, True),
     cfgv.Optional('description', cfgv.check_string, ''),
     cfgv.Optional('language_version', cfgv.check_string, C.DEFAULT),
     cfgv.Optional('log_file', cfgv.check_string, ''),
+    cfgv.Optional('minimum_pre_commit_version', cfgv.check_string, '0'),
     cfgv.Optional('require_serial', cfgv.check_bool, False),
-    StagesMigration('stages', []),
+    cfgv.Optional('stages', cfgv.check_array(cfgv.check_one_of(C.STAGES)), []),
     cfgv.Optional('verbose', cfgv.check_bool, False),
 )
 MANIFEST_SCHEMA = cfgv.Array(MANIFEST_HOOK_DICT)
@@ -270,29 +86,36 @@ class InvalidManifestError(FatalError):
     pass
 
 
-def _load_manifest_forward_compat(contents: str) -> object:
-    obj = yaml_load(contents)
-    if isinstance(obj, dict):
-        check_min_version('5')
-        raise AssertionError('unreachable')
-    else:
-        return obj
-
-
 load_manifest = functools.partial(
     cfgv.load_from_filename,
     schema=MANIFEST_SCHEMA,
-    load_strategy=_load_manifest_forward_compat,
+    load_strategy=yaml_load,
     exc_tp=InvalidManifestError,
 )
 
 
+def validate_manifest_main(argv: Optional[Sequence[str]] = None) -> int:
+    parser = _make_argparser('Manifest filenames.')
+    args = parser.parse_args(argv)
+
+    with logging_handler(args.color):
+        ret = 0
+        for filename in args.filenames:
+            try:
+                load_manifest(filename)
+            except InvalidManifestError as e:
+                print(e)
+                ret = 1
+        return ret
+
+
 LOCAL = 'local'
 META = 'meta'
 
 
-class WarnMutableRev(cfgv.Conditional):
-    def check(self, dct: dict[str, Any]) -> None:
+# should inherit from cfgv.Conditional if sha support is dropped
+class WarnMutableRev(cfgv.ConditionalOptional):
+    def check(self, dct: Dict[str, Any]) -> None:
         super().check(dct)
 
         if self.key in dct:
@@ -311,7 +134,7 @@ class WarnMutableRev(cfgv.Conditional):
 
 
 class OptionalSensibleRegexAtHook(cfgv.OptionalNoDefault):
-    def check(self, dct: dict[str, Any]) -> None:
+    def check(self, dct: Dict[str, Any]) -> None:
         super().check(dct)
 
         if '/*' in dct.get(self.key, ''):
@@ -320,17 +143,10 @@ class OptionalSensibleRegexAtHook(cfgv.OptionalNoDefault):
                 f"regex, not a glob -- matching '/*' probably isn't what you "
                 f'want here',
             )
-        for fwd_slash_re in (r'[\\/]', r'[\/]', r'[/\\]'):
-            if fwd_slash_re in dct.get(self.key, ''):
-                logger.warning(
-                    fr'pre-commit normalizes slashes in the {self.key!r} '
-                    fr'field in hook {dct.get("id")!r} to forward slashes, '
-                    fr'so you can use / instead of {fwd_slash_re}',
-                )
 
 
 class OptionalSensibleRegexAtTop(cfgv.OptionalNoDefault):
-    def check(self, dct: dict[str, Any]) -> None:
+    def check(self, dct: Dict[str, Any]) -> None:
         super().check(dct)
 
         if '/*' in dct.get(self.key, ''):
@@ -338,13 +154,36 @@ class OptionalSensibleRegexAtTop(cfgv.OptionalNoDefault):
                 f'The top-level {self.key!r} field is a regex, not a glob -- '
                 f"matching '/*' probably isn't what you want here",
             )
-        for fwd_slash_re in (r'[\\/]', r'[\/]', r'[/\\]'):
-            if fwd_slash_re in dct.get(self.key, ''):
-                logger.warning(
-                    fr'pre-commit normalizes the slashes in the top-level '
-                    fr'{self.key!r} field to forward slashes, so you '
-                    fr'can use / instead of {fwd_slash_re}',
-                )
+
+
+class MigrateShaToRev:
+    key = 'rev'
+
+    @staticmethod
+    def _cond(key: str) -> cfgv.Conditional:
+        return cfgv.Conditional(
+            key, cfgv.check_string,
+            condition_key='repo',
+            condition_value=cfgv.NotIn(LOCAL, META),
+            ensure_absent=True,
+        )
+
+    def check(self, dct: Dict[str, Any]) -> None:
+        if dct.get('repo') in {LOCAL, META}:
+            self._cond('rev').check(dct)
+            self._cond('sha').check(dct)
+        elif 'sha' in dct and 'rev' in dct:
+            raise cfgv.ValidationError('Cannot specify both sha and rev')
+        elif 'sha' in dct:
+            self._cond('sha').check(dct)
+        else:
+            self._cond('rev').check(dct)
+
+    def apply_default(self, dct: Dict[str, Any]) -> None:
+        if 'sha' in dct:
+            dct['rev'] = dct.pop('sha')
+
+    remove_default = cfgv.Required.remove_default
 
 
 def _entry(modname: str) -> str:
@@ -358,7 +197,7 @@ def _entry(modname: str) -> str:
 def warn_unknown_keys_root(
         extra: Sequence[str],
         orig_keys: Sequence[str],
-        dct: dict[str, str],
+        dct: Dict[str, str],
 ) -> None:
     logger.warning(f'Unexpected key(s) present at root: {", ".join(extra)}')
 
@@ -366,7 +205,7 @@ def warn_unknown_keys_root(
 def warn_unknown_keys_repo(
         extra: Sequence[str],
         orig_keys: Sequence[str],
-        dct: dict[str, str],
+        dct: Dict[str, str],
 ) -> None:
     logger.warning(
         f'Unexpected key(s) present on {dct["repo"]}: {", ".join(extra)}',
@@ -397,29 +236,12 @@ _meta = (
     ),
 )
 
-
-class NotAllowed(cfgv.OptionalNoDefault):
-    def check(self, dct: dict[str, Any]) -> None:
-        if self.key in dct:
-            raise cfgv.ValidationError(f'{self.key!r} cannot be overridden')
-
-
-_COMMON_HOOK_WARNINGS = (
-    OptionalSensibleRegexAtHook('files', cfgv.check_string),
-    OptionalSensibleRegexAtHook('exclude', cfgv.check_string),
-    DeprecatedStagesWarning('stages'),
-)
-
 META_HOOK_DICT = cfgv.Map(
     'Hook', 'id',
     cfgv.Required('id', cfgv.check_string),
     cfgv.Required('id', cfgv.check_one_of(tuple(k for k, _ in _meta))),
-    # language must be `unsupported`
-    cfgv.Optional(
-        'language', cfgv.check_one_of({'unsupported'}), 'unsupported',
-    ),
-    # entry cannot be overridden
-    NotAllowed('entry', cfgv.check_any),
+    # language must be system
+    cfgv.Optional('language', cfgv.check_one_of({'system'}), 'system'),
     *(
         # default to the hook definition for the meta hooks
         cfgv.ConditionalOptional(key, cfgv.check_any, value, 'id', hook_id)
@@ -434,7 +256,6 @@ META_HOOK_DICT = cfgv.Map(
         item
         for item in MANIFEST_HOOK_DICT.items
     ),
-    *_COMMON_HOOK_WARNINGS,
 )
 CONFIG_HOOK_DICT = cfgv.Map(
     'Hook', 'id',
@@ -449,18 +270,9 @@ CONFIG_HOOK_DICT = cfgv.Map(
         cfgv.OptionalNoDefault(item.key, item.check_fn)
         for item in MANIFEST_HOOK_DICT.items
         if item.key != 'id'
-        if item.key != 'stages'
-        if item.key != 'language'  # remove
     ),
-    StagesMigrationNoDefault('stages', []),
-    LanguageMigration('language', cfgv.check_one_of(language_names)),  # remove
-    *_COMMON_HOOK_WARNINGS,
-)
-LOCAL_HOOK_DICT = cfgv.Map(
-    'Hook', 'id',
-
-    *MANIFEST_HOOK_DICT.items,
-    *_COMMON_HOOK_WARNINGS,
+    OptionalSensibleRegexAtHook('files', cfgv.check_string),
+    OptionalSensibleRegexAtHook('exclude', cfgv.check_string),
 )
 CONFIG_REPO_DICT = cfgv.Map(
     'Repository', 'repo',
@@ -472,7 +284,7 @@ CONFIG_REPO_DICT = cfgv.Map(
         'repo', cfgv.NotIn(LOCAL, META),
     ),
     cfgv.ConditionalRecurse(
-        'hooks', cfgv.Array(LOCAL_HOOK_DICT),
+        'hooks', cfgv.Array(MANIFEST_HOOK_DICT),
         'repo', LOCAL,
     ),
     cfgv.ConditionalRecurse(
@@ -480,47 +292,45 @@ CONFIG_REPO_DICT = cfgv.Map(
         'repo', META,
     ),
 
+    MigrateShaToRev(),
     WarnMutableRev(
-        'rev', cfgv.check_string,
-        condition_key='repo',
-        condition_value=cfgv.NotIn(LOCAL, META),
-        ensure_absent=True,
+        'rev',
+        cfgv.check_string,
+        '',
+        'repo',
+        cfgv.NotIn(LOCAL, META),
+        True,
     ),
     cfgv.WarnAdditionalKeys(('repo', 'rev', 'hooks'), warn_unknown_keys_repo),
 )
 DEFAULT_LANGUAGE_VERSION = cfgv.Map(
     'DefaultLanguageVersion', None,
-    cfgv.NoAdditionalKeys(language_names),
-    *(cfgv.Optional(x, cfgv.check_string, C.DEFAULT) for x in language_names),
+    cfgv.NoAdditionalKeys(all_languages),
+    *(cfgv.Optional(x, cfgv.check_string, C.DEFAULT) for x in all_languages),
 )
 CONFIG_SCHEMA = cfgv.Map(
     'Config', None,
 
-    # check first in case it uses some newer, incompatible feature
+    cfgv.RequiredRecurse('repos', cfgv.Array(CONFIG_REPO_DICT)),
+    cfgv.OptionalRecurse(
+        'default_language_version', DEFAULT_LANGUAGE_VERSION, {},
+    ),
+    cfgv.Optional(
+        'default_stages',
+        cfgv.check_array(cfgv.check_one_of(C.STAGES)),
+        C.STAGES,
+    ),
+    cfgv.Optional('files', check_string_regex, ''),
+    cfgv.Optional('exclude', check_string_regex, '^$'),
+    cfgv.Optional('fail_fast', cfgv.check_bool, False),
     cfgv.Optional(
         'minimum_pre_commit_version',
         cfgv.check_and(cfgv.check_string, check_min_version),
         '0',
     ),
-
-    cfgv.RequiredRecurse('repos', cfgv.Array(CONFIG_REPO_DICT)),
-    cfgv.Optional(
-        'default_install_hook_types',
-        cfgv.check_array(cfgv.check_one_of(HOOK_TYPES)),
-        ['pre-commit'],
-    ),
-    cfgv.OptionalRecurse(
-        'default_language_version', DEFAULT_LANGUAGE_VERSION, {},
-    ),
-    StagesMigration('default_stages', STAGES),
-    DeprecatedDefaultStagesWarning('default_stages'),
-    cfgv.Optional('files', check_string_regex, ''),
-    cfgv.Optional('exclude', check_string_regex, '^$'),
-    cfgv.Optional('fail_fast', cfgv.check_bool, False),
     cfgv.WarnAdditionalKeys(
         (
             'repos',
-            'default_install_hook_types',
             'default_language_version',
             'default_stages',
             'files',
@@ -543,9 +353,37 @@ class InvalidConfigError(FatalError):
     pass
 
 
+def ordered_load_normalize_legacy_config(contents: str) -> Dict[str, Any]:
+    data = yaml_load(contents)
+    if isinstance(data, list):
+        logger.warning(
+            'normalizing pre-commit configuration to a top-level map.  '
+            'support for top level list will be removed in a future version.  '
+            'run: `pre-commit migrate-config` to automatically fix this.',
+        )
+        return {'repos': data}
+    else:
+        return data
+
+
 load_config = functools.partial(
     cfgv.load_from_filename,
     schema=CONFIG_SCHEMA,
-    load_strategy=yaml_load,
+    load_strategy=ordered_load_normalize_legacy_config,
     exc_tp=InvalidConfigError,
 )
+
+
+def validate_config_main(argv: Optional[Sequence[str]] = None) -> int:
+    parser = _make_argparser('Config filenames.')
+    args = parser.parse_args(argv)
+
+    with logging_handler(args.color):
+        ret = 0
+        for filename in args.filenames:
+            try:
+                load_config(filename)
+            except InvalidConfigError as e:
+                print(e)
+                ret = 1
+        return ret

pre_commit/color.py

@@ -1,5 +1,3 @@
-from __future__ import annotations
-
 import argparse
 import os
 import sys

pre_commit/commands/autoupdate.py

@@ -1,85 +1,62 @@
-from __future__ import annotations
-
-import concurrent.futures
 import os.path
 import re
-import tempfile
-from collections.abc import Sequence
 from typing import Any
+from typing import Dict
+from typing import List
 from typing import NamedTuple
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
 
 import pre_commit.constants as C
 from pre_commit import git
 from pre_commit import output
-from pre_commit import xargs
 from pre_commit.clientlib import InvalidManifestError
 from pre_commit.clientlib import load_config
 from pre_commit.clientlib import load_manifest
 from pre_commit.clientlib import LOCAL
 from pre_commit.clientlib import META
 from pre_commit.commands.migrate_config import migrate_config
+from pre_commit.store import Store
 from pre_commit.util import CalledProcessError
 from pre_commit.util import cmd_output
 from pre_commit.util import cmd_output_b
-from pre_commit.yaml import yaml_dump
-from pre_commit.yaml import yaml_load
+from pre_commit.util import tmpdir
+from pre_commit.util import yaml_dump
+from pre_commit.util import yaml_load
 
 
 class RevInfo(NamedTuple):
     repo: str
     rev: str
-    frozen: str | None = None
-    hook_ids: frozenset[str] = frozenset()
+    frozen: Optional[str]
 
     @classmethod
-    def from_config(cls, config: dict[str, Any]) -> RevInfo:
-        return cls(config['repo'], config['rev'])
+    def from_config(cls, config: Dict[str, Any]) -> 'RevInfo':
+        return cls(config['repo'], config['rev'], None)
 
-    def update(self, tags_only: bool, freeze: bool) -> RevInfo:
-        with tempfile.TemporaryDirectory() as tmp:
-            _git = ('git', *git.NO_FS_MONITOR, '-C', tmp)
-
-            if tags_only:
-                tag_opt = '--abbrev=0'
-            else:
-                tag_opt = '--exact'
-            tag_cmd = (*_git, 'describe', 'FETCH_HEAD', '--tags', tag_opt)
+    def update(self, tags_only: bool, freeze: bool) -> 'RevInfo':
+        if tags_only:
+            tag_cmd = ('git', 'describe', 'FETCH_HEAD', '--tags', '--abbrev=0')
+        else:
+            tag_cmd = ('git', 'describe', 'FETCH_HEAD', '--tags', '--exact')
 
+        with tmpdir() as tmp:
             git.init_repo(tmp, self.repo)
-            cmd_output_b(*_git, 'config', 'extensions.partialClone', 'true')
-            cmd_output_b(
-                *_git, 'fetch', 'origin', 'HEAD',
-                '--quiet', '--filter=blob:none', '--tags',
-            )
+            cmd_output_b('git', 'fetch', 'origin', 'HEAD', '--tags', cwd=tmp)
 
             try:
-                rev = cmd_output(*tag_cmd)[1].strip()
+                rev = cmd_output(*tag_cmd, cwd=tmp)[1].strip()
             except CalledProcessError:
-                rev = cmd_output(*_git, 'rev-parse', 'FETCH_HEAD')[1].strip()
-            else:
-                if tags_only:
-                    rev = git.get_best_candidate_tag(rev, tmp)
+                cmd = ('git', 'rev-parse', 'FETCH_HEAD')
+                rev = cmd_output(*cmd, cwd=tmp)[1].strip()
 
             frozen = None
             if freeze:
-                exact = cmd_output(*_git, 'rev-parse', rev)[1].strip()
+                exact = cmd_output('git', 'rev-parse', rev, cwd=tmp)[1].strip()
                 if exact != rev:
                     rev, frozen = exact, rev
-
-            try:
-                # workaround for windows -- see #2865
-                cmd_output_b(*_git, 'show', f'{rev}:{C.MANIFEST_FILE}')
-                cmd_output(*_git, 'checkout', rev, '--', C.MANIFEST_FILE)
-            except CalledProcessError:
-                pass  # this will be caught by manifest validating code
-            try:
-                manifest = load_manifest(os.path.join(tmp, C.MANIFEST_FILE))
-            except InvalidManifestError as e:
-                raise RepositoryCannotBeUpdatedError(f'[{self.repo}] {e}')
-            else:
-                hook_ids = frozenset(hook['id'] for hook in manifest)
-
-        return self._replace(rev=rev, frozen=frozen, hook_ids=hook_ids)
+        return self._replace(rev=rev, frozen=frozen)
 
 
 class RepositoryCannotBeUpdatedError(RuntimeError):
@@ -87,40 +64,34 @@ class RepositoryCannotBeUpdatedError(RuntimeError):
 
 
 def _check_hooks_still_exist_at_rev(
-        repo_config: dict[str, Any],
+        repo_config: Dict[str, Any],
         info: RevInfo,
+        store: Store,
 ) -> None:
+    try:
+        path = store.clone(repo_config['repo'], info.rev)
+        manifest = load_manifest(os.path.join(path, C.MANIFEST_FILE))
+    except InvalidManifestError as e:
+        raise RepositoryCannotBeUpdatedError(str(e))
+
     # See if any of our hooks were deleted with the new commits
     hooks = {hook['id'] for hook in repo_config['hooks']}
-    hooks_missing = hooks - info.hook_ids
+    hooks_missing = hooks - {hook['id'] for hook in manifest}
     if hooks_missing:
         raise RepositoryCannotBeUpdatedError(
-            f'[{info.repo}] Cannot update because the update target is '
-            f'missing these hooks: {", ".join(sorted(hooks_missing))}',
+            f'Cannot update because the update target is missing these '
+            f'hooks:\n{", ".join(sorted(hooks_missing))}',
         )
 
 
-def _update_one(
-        i: int,
-        repo: dict[str, Any],
-        *,
-        tags_only: bool,
-        freeze: bool,
-) -> tuple[int, RevInfo, RevInfo]:
-    old = RevInfo.from_config(repo)
-    new = old.update(tags_only=tags_only, freeze=freeze)
-    _check_hooks_still_exist_at_rev(repo, new)
-    return i, old, new
-
-
 REV_LINE_RE = re.compile(r'^(\s+)rev:(\s*)([\'"]?)([^\s#]+)(.*)(\r?\n)$')
 
 
 def _original_lines(
         path: str,
-        rev_infos: list[RevInfo | None],
+        rev_infos: List[Optional[RevInfo]],
         retry: bool = False,
-) -> tuple[list[str], list[int]]:
+) -> Tuple[List[str], List[int]]:
     """detect `rev:` lines or reformat the file"""
     with open(path, newline='') as f:
         original = f.read()
@@ -137,7 +108,7 @@ def _original_lines(
         return _original_lines(path, rev_infos, retry=True)
 
 
-def _write_new_config(path: str, rev_infos: list[RevInfo | None]) -> None:
+def _write_new_config(path: str, rev_infos: List[Optional[RevInfo]]) -> None:
     lines, idxs = _original_lines(path, rev_infos)
 
     for idx, rev_info in zip(idxs, rev_infos):
@@ -161,53 +132,49 @@ def _write_new_config(path: str, rev_infos: list[RevInfo | None]) -> None:
 
 def autoupdate(
         config_file: str,
+        store: Store,
         tags_only: bool,
         freeze: bool,
         repos: Sequence[str] = (),
-        jobs: int = 1,
 ) -> int:
     """Auto-update the pre-commit config to the latest versions of repos."""
     migrate_config(config_file, quiet=True)
-    changed = False
     retv = 0
+    rev_infos: List[Optional[RevInfo]] = []
+    changed = False
 
-    config_repos = [
-        repo for repo in load_config(config_file)['repos']
-        if repo['repo'] not in {LOCAL, META}
-    ]
+    config = load_config(config_file)
+    for repo_config in config['repos']:
+        if repo_config['repo'] in {LOCAL, META}:
+            continue
 
-    rev_infos: list[RevInfo | None] = [None] * len(config_repos)
-    jobs = jobs or xargs.cpu_count()  # 0 => number of cpus
-    jobs = min(jobs, len(repos) or len(config_repos))  # max 1-per-thread
-    jobs = max(jobs, 1)  # at least one thread
-    with concurrent.futures.ThreadPoolExecutor(jobs) as exe:
-        futures = [
-            exe.submit(
-                _update_one,
-                i, repo, tags_only=tags_only, freeze=freeze,
-            )
-            for i, repo in enumerate(config_repos)
-            if not repos or repo['repo'] in repos
-        ]
-        for future in concurrent.futures.as_completed(futures):
-            try:
-                i, old, new = future.result()
-            except RepositoryCannotBeUpdatedError as e:
-                output.write_line(str(e))
-                retv = 1
+        info = RevInfo.from_config(repo_config)
+        if repos and info.repo not in repos:
+            rev_infos.append(None)
+            continue
+
+        output.write(f'Updating {info.repo} ... ')
+        new_info = info.update(tags_only=tags_only, freeze=freeze)
+        try:
+            _check_hooks_still_exist_at_rev(repo_config, new_info, store)
+        except RepositoryCannotBeUpdatedError as error:
+            output.write_line(error.args[0])
+            rev_infos.append(None)
+            retv = 1
+            continue
+
+        if new_info.rev != info.rev:
+            changed = True
+            if new_info.frozen:
+                updated_to = f'{new_info.frozen} (frozen)'
             else:
-                if new.rev != old.rev:
-                    changed = True
-                    if new.frozen:
-                        new_s = f'{new.frozen} (frozen)'
-                    else:
-                        new_s = new.rev
-                    msg = f'updating {old.rev} -> {new_s}'
-                    rev_infos[i] = new
-                else:
-                    msg = 'already up to date!'
-
-                output.write_line(f'[{old.repo}] {msg}')
+                updated_to = new_info.rev
+            msg = f'updating {info.rev} -> {updated_to}.'
+            output.write_line(msg)
+            rev_infos.append(new_info)
+        else:
+            output.write_line('already up to date.')
+            rev_infos.append(None)
 
     if changed:
         _write_new_config(config_file, rev_infos)

pre_commit/commands/clean.py

@@ -1,5 +1,3 @@
-from __future__ import annotations
-
 import os.path
 
 from pre_commit import output

pre_commit/commands/gc.py

@@ -1,7 +1,8 @@
-from __future__ import annotations
-
 import os.path
 from typing import Any
+from typing import Dict
+from typing import Set
+from typing import Tuple
 
 import pre_commit.constants as C
 from pre_commit import output
@@ -12,14 +13,13 @@ from pre_commit.clientlib import load_manifest
 from pre_commit.clientlib import LOCAL
 from pre_commit.clientlib import META
 from pre_commit.store import Store
-from pre_commit.util import rmtree
 
 
 def _mark_used_repos(
         store: Store,
-        all_repos: dict[tuple[str, str], str],
-        unused_repos: set[tuple[str, str]],
-        repo: dict[str, Any],
+        all_repos: Dict[Tuple[str, str], str],
+        unused_repos: Set[Tuple[str, str]],
+        repo: Dict[str, Any],
 ) -> None:
     if repo['repo'] == META:
         return
@@ -27,8 +27,7 @@ def _mark_used_repos(
         for hook in repo['hooks']:
             deps = hook.get('additional_dependencies')
             unused_repos.discard((
-                store.db_repo_name(repo['repo'], deps),
-                C.LOCAL_REPO_VERSION,
+                store.db_repo_name(repo['repo'], deps), C.LOCAL_REPO_VERSION,
             ))
     else:
         key = (repo['repo'], repo['rev'])
@@ -58,41 +57,34 @@ def _mark_used_repos(
             ))
 
 
-def _gc(store: Store) -> int:
-    with store.exclusive_lock(), store.connect() as db:
-        store._create_configs_table(db)
+def _gc_repos(store: Store) -> int:
+    configs = store.select_all_configs()
+    repos = store.select_all_repos()
 
-        repos = db.execute('SELECT repo, ref, path FROM repos').fetchall()
-        all_repos = {(repo, ref): path for repo, ref, path in repos}
-        unused_repos = set(all_repos)
+    # delete config paths which do not exist
+    dead_configs = [p for p in configs if not os.path.exists(p)]
+    live_configs = [p for p in configs if os.path.exists(p)]
 
-        configs_rows = db.execute('SELECT path FROM configs').fetchall()
-        configs = [path for path, in configs_rows]
+    all_repos = {(repo, ref): path for repo, ref, path in repos}
+    unused_repos = set(all_repos)
+    for config_path in live_configs:
+        try:
+            config = load_config(config_path)
+        except InvalidConfigError:
+            dead_configs.append(config_path)
+            continue
+        else:
+            for repo in config['repos']:
+                _mark_used_repos(store, all_repos, unused_repos, repo)
 
-        dead_configs = []
-        for config_path in configs:
-            try:
-                config = load_config(config_path)
-            except InvalidConfigError:
-                dead_configs.append(config_path)
-                continue
-            else:
-                for repo in config['repos']:
-                    _mark_used_repos(store, all_repos, unused_repos, repo)
-
-        paths = [(path,) for path in dead_configs]
-        db.executemany('DELETE FROM configs WHERE path = ?', paths)
-
-        db.executemany(
-            'DELETE FROM repos WHERE repo = ? and ref = ?',
-            sorted(unused_repos),
-        )
-        for k in unused_repos:
-            rmtree(all_repos[k])
-
-        return len(unused_repos)
+    store.delete_configs(dead_configs)
+    for db_repo_name, ref in unused_repos:
+        store.delete_repo(db_repo_name, ref, all_repos[(db_repo_name, ref)])
+    return len(unused_repos)
 
 
 def gc(store: Store) -> int:
-    output.write_line(f'{_gc(store)} repo(s) removed.')
+    with store.exclusive_lock():
+        repos_removed = _gc_repos(store)
+    output.write_line(f'{repos_removed} repo(s) removed.')
     return 0

pre_commit/commands/hazmat.py

@@ -1,95 +0,0 @@
-from __future__ import annotations
-
-import argparse
-import subprocess
-from collections.abc import Sequence
-
-from pre_commit.parse_shebang import normalize_cmd
-
-
-def add_parsers(parser: argparse.ArgumentParser) -> None:
-    subparsers = parser.add_subparsers(dest='tool')
-
-    cd_parser = subparsers.add_parser(
-        'cd', help='cd to a subdir and run the command',
-    )
-    cd_parser.add_argument('subdir')
-    cd_parser.add_argument('cmd', nargs=argparse.REMAINDER)
-
-    ignore_exit_code_parser = subparsers.add_parser(
-        'ignore-exit-code', help='run the command but ignore the exit code',
-    )
-    ignore_exit_code_parser.add_argument('cmd', nargs=argparse.REMAINDER)
-
-    n1_parser = subparsers.add_parser(
-        'n1', help='run the command once per filename',
-    )
-    n1_parser.add_argument('cmd', nargs=argparse.REMAINDER)
-
-
-def _cmd_filenames(cmd: tuple[str, ...]) -> tuple[
-    tuple[str, ...],
-    tuple[str, ...],
-]:
-    for idx, val in enumerate(reversed(cmd)):
-        if val == '--':
-            split = len(cmd) - idx
-            break
-    else:
-        raise SystemExit('hazmat entry must end with `--`')
-
-    return cmd[:split - 1], cmd[split:]
-
-
-def cd(subdir: str, cmd: tuple[str, ...]) -> int:
-    cmd, filenames = _cmd_filenames(cmd)
-
-    prefix = f'{subdir}/'
-    new_filenames = []
-    for filename in filenames:
-        if not filename.startswith(prefix):
-            raise SystemExit(f'unexpected file without {prefix=}: {filename}')
-        else:
-            new_filenames.append(filename.removeprefix(prefix))
-
-    cmd = normalize_cmd(cmd)
-    return subprocess.call((*cmd, *new_filenames), cwd=subdir)
-
-
-def ignore_exit_code(cmd: tuple[str, ...]) -> int:
-    cmd = normalize_cmd(cmd)
-    subprocess.call(cmd)
-    return 0
-
-
-def n1(cmd: tuple[str, ...]) -> int:
-    cmd, filenames = _cmd_filenames(cmd)
-    cmd = normalize_cmd(cmd)
-    ret = 0
-    for filename in filenames:
-        ret |= subprocess.call((*cmd, filename))
-    return ret
-
-
-def impl(args: argparse.Namespace) -> int:
-    args.cmd = tuple(args.cmd)
-    if args.tool == 'cd':
-        return cd(args.subdir, args.cmd)
-    elif args.tool == 'ignore-exit-code':
-        return ignore_exit_code(args.cmd)
-    elif args.tool == 'n1':
-        return n1(args.cmd)
-    else:
-        raise NotImplementedError(f'unexpected tool: {args.tool}')
-
-
-def main(argv: Sequence[str] | None = None) -> int:
-    parser = argparse.ArgumentParser()
-    add_parsers(parser)
-    args = parser.parse_args(argv)
-
-    return impl(args)
-
-
-if __name__ == '__main__':
-    raise SystemExit(main())

pre_commit/commands/hook_impl.py

@@ -1,10 +1,10 @@
-from __future__ import annotations
-
 import argparse
 import os.path
 import subprocess
 import sys
-from collections.abc import Sequence
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
 
 from pre_commit.commands.run import run
 from pre_commit.envcontext import envcontext
@@ -18,7 +18,7 @@ def _run_legacy(
         hook_type: str,
         hook_dir: str,
         args: Sequence[str],
-) -> tuple[int, bytes]:
+) -> Tuple[int, bytes]:
     if os.environ.get('PRE_COMMIT_RUNNING_LEGACY'):
         raise SystemExit(
             f"bug: pre-commit's script is installed in migration mode\n"
@@ -69,35 +69,27 @@ def _ns(
         color: bool,
         *,
         all_files: bool = False,
-        remote_branch: str | None = None,
-        local_branch: str | None = None,
-        from_ref: str | None = None,
-        to_ref: str | None = None,
-        pre_rebase_upstream: str | None = None,
-        pre_rebase_branch: str | None = None,
-        remote_name: str | None = None,
-        remote_url: str | None = None,
-        commit_msg_filename: str | None = None,
-        prepare_commit_message_source: str | None = None,
-        commit_object_name: str | None = None,
-        checkout_type: str | None = None,
-        is_squash_merge: str | None = None,
-        rewrite_command: str | None = None,
+        remote_branch: Optional[str] = None,
+        local_branch: Optional[str] = None,
+        from_ref: Optional[str] = None,
+        to_ref: Optional[str] = None,
+        remote_name: Optional[str] = None,
+        remote_url: Optional[str] = None,
+        commit_msg_filename: Optional[str] = None,
+        checkout_type: Optional[str] = None,
+        is_squash_merge: Optional[str] = None,
+        rewrite_command: Optional[str] = None,
 ) -> argparse.Namespace:
     return argparse.Namespace(
         color=color,
-        hook_stage=hook_type,
+        hook_stage=hook_type.replace('pre-', ''),
         remote_branch=remote_branch,
         local_branch=local_branch,
         from_ref=from_ref,
         to_ref=to_ref,
-        pre_rebase_upstream=pre_rebase_upstream,
-        pre_rebase_branch=pre_rebase_branch,
         remote_name=remote_name,
         remote_url=remote_url,
         commit_msg_filename=commit_msg_filename,
-        prepare_commit_message_source=prepare_commit_message_source,
-        commit_object_name=commit_object_name,
         all_files=all_files,
         checkout_type=checkout_type,
         is_squash_merge=is_squash_merge,
@@ -106,7 +98,6 @@ def _ns(
         hook=None,
         verbose=False,
         show_diff_on_failure=False,
-        fail_fast=False,
     )
 
 
@@ -118,13 +109,12 @@ def _pre_push_ns(
         color: bool,
         args: Sequence[str],
         stdin: bytes,
-) -> argparse.Namespace | None:
+) -> Optional[argparse.Namespace]:
     remote_name = args[0]
     remote_url = args[1]
 
     for line in stdin.decode().splitlines():
-        parts = line.rsplit(maxsplit=3)
-        local_branch, local_sha, remote_branch, remote_sha = parts
+        local_branch, local_sha, remote_branch, remote_sha = line.split()
         if local_sha == Z40:
             continue
         elif remote_sha != Z40 and _rev_exists(remote_sha):
@@ -190,12 +180,6 @@ def _check_args_length(hook_type: str, args: Sequence[str]) -> None:
                 f'hook-impl for {hook_type} expected 1, 2, or 3 arguments '
                 f'but got {len(args)}: {args}',
             )
-    elif hook_type == 'pre-rebase':
-        if len(args) < 1 or len(args) > 2:
-            raise SystemExit(
-                f'hook-impl for {hook_type} expected 1 or 2 arguments '
-                f'but got {len(args)}: {args}',
-            )
     elif hook_type in _EXPECTED_ARG_LENGTH_BY_HOOK:
         expected = _EXPECTED_ARG_LENGTH_BY_HOOK[hook_type]
         if len(args) != expected:
@@ -213,24 +197,12 @@ def _run_ns(
         color: bool,
         args: Sequence[str],
         stdin: bytes,
-) -> argparse.Namespace | None:
+) -> Optional[argparse.Namespace]:
     _check_args_length(hook_type, args)
     if hook_type == 'pre-push':
         return _pre_push_ns(color, args, stdin)
-    elif hook_type in 'commit-msg':
+    elif hook_type in {'commit-msg', 'prepare-commit-msg'}:
         return _ns(hook_type, color, commit_msg_filename=args[0])
-    elif hook_type == 'prepare-commit-msg' and len(args) == 1:
-        return _ns(hook_type, color, commit_msg_filename=args[0])
-    elif hook_type == 'prepare-commit-msg' and len(args) == 2:
-        return _ns(
-            hook_type, color, commit_msg_filename=args[0],
-            prepare_commit_message_source=args[1],
-        )
-    elif hook_type == 'prepare-commit-msg' and len(args) == 3:
-        return _ns(
-            hook_type, color, commit_msg_filename=args[0],
-            prepare_commit_message_source=args[1], commit_object_name=args[2],
-        )
     elif hook_type in {'post-commit', 'pre-merge-commit', 'pre-commit'}:
         return _ns(hook_type, color)
     elif hook_type == 'post-checkout':
@@ -242,13 +214,6 @@ def _run_ns(
         return _ns(hook_type, color, is_squash_merge=args[0])
     elif hook_type == 'post-rewrite':
         return _ns(hook_type, color, rewrite_command=args[0])
-    elif hook_type == 'pre-rebase' and len(args) == 1:
-        return _ns(hook_type, color, pre_rebase_upstream=args[0])
-    elif hook_type == 'pre-rebase' and len(args) == 2:
-        return _ns(
-            hook_type, color, pre_rebase_upstream=args[0],
-            pre_rebase_branch=args[1],
-        )
     else:
         raise AssertionError(f'unexpected hook type: {hook_type}')
 

pre_commit/commands/init_templatedir.py

@@ -1,7 +1,6 @@
-from __future__ import annotations
-
 import logging
 import os.path
+from typing import Sequence
 
 from pre_commit.commands.install_uninstall import install
 from pre_commit.store import Store
@@ -15,7 +14,7 @@ def init_templatedir(
         config_file: str,
         store: Store,
         directory: str,
-        hook_types: list[str] | None,
+        hook_types: Sequence[str],
         skip_on_missing_config: bool = True,
 ) -> int:
     install(

pre_commit/commands/install_uninstall.py

@@ -1,14 +1,14 @@
-from __future__ import annotations
-
+import itertools
 import logging
 import os.path
-import shlex
 import shutil
 import sys
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
 
 from pre_commit import git
 from pre_commit import output
-from pre_commit.clientlib import InvalidConfigError
 from pre_commit.clientlib import load_config
 from pre_commit.repository import all_hooks
 from pre_commit.repository import install_hook_envs
@@ -30,25 +30,17 @@ PRIOR_HASHES = (
 CURRENT_HASH = b'138fd403232d2ddd5efb44317e38bf03'
 TEMPLATE_START = '# start templated\n'
 TEMPLATE_END = '# end templated\n'
-
-
-def _hook_types(cfg_filename: str, hook_types: list[str] | None) -> list[str]:
-    if hook_types is not None:
-        return hook_types
-    else:
-        try:
-            cfg = load_config(cfg_filename)
-        except InvalidConfigError:
-            return ['pre-commit']
-        else:
-            return cfg['default_install_hook_types']
+# Homebrew/homebrew-core#35825: be more timid about appropriate `PATH`
+# #1312 os.defpath is too restrictive on BSD
+POSIX_SEARCH_PATH = ('/usr/local/bin', '/usr/bin', '/bin')
+SYS_EXE = os.path.basename(os.path.realpath(sys.executable))
 
 
 def _hook_paths(
         hook_type: str,
-        git_dir: str | None = None,
-) -> tuple[str, str]:
-    git_dir = git_dir if git_dir is not None else git.get_git_common_dir()
+        git_dir: Optional[str] = None,
+) -> Tuple[str, str]:
+    git_dir = git_dir if git_dir is not None else git.get_git_dir()
     pth = os.path.join(git_dir, 'hooks', hook_type)
     return pth, f'{pth}.legacy'
 
@@ -61,12 +53,32 @@ def is_our_script(filename: str) -> bool:
     return any(h in contents for h in (CURRENT_HASH,) + PRIOR_HASHES)
 
 
+def shebang() -> str:
+    if sys.platform == 'win32':
+        py, _ = os.path.splitext(SYS_EXE)
+    else:
+        exe_choices = [
+            f'python{sys.version_info[0]}.{sys.version_info[1]}',
+            f'python{sys.version_info[0]}',
+        ]
+        # avoid searching for bare `python` as it's likely to be python 2
+        if SYS_EXE != 'python':
+            exe_choices.append(SYS_EXE)
+        for path, exe in itertools.product(POSIX_SEARCH_PATH, exe_choices):
+            if os.access(os.path.join(path, exe), os.X_OK):
+                py = exe
+                break
+        else:
+            py = SYS_EXE
+    return f'#!/usr/bin/env {py}'
+
+
 def _install_hook_script(
         config_file: str,
         hook_type: str,
         overwrite: bool = False,
         skip_on_missing_config: bool = False,
-        git_dir: str | None = None,
+        git_dir: Optional[str] = None,
 ) -> None:
     hook_path, legacy_path = _hook_paths(hook_type, git_dir=git_dir)
 
@@ -88,23 +100,19 @@ def _install_hook_script(
     args = ['hook-impl', f'--config={config_file}', f'--hook-type={hook_type}']
     if skip_on_missing_config:
         args.append('--skip-on-missing-config')
+    params = {'INSTALL_PYTHON': sys.executable, 'ARGS': args}
 
     with open(hook_path, 'w') as hook_file:
         contents = resource_text('hook-tmpl')
         before, rest = contents.split(TEMPLATE_START)
-        _, after = rest.split(TEMPLATE_END)
+        to_template, after = rest.split(TEMPLATE_END)
 
-        # on windows always use `/bin/sh` since `bash` might not be on PATH
-        # though we use bash-specific features `sh` on windows is actually
-        # bash in "POSIXLY_CORRECT" mode which still supports the features we
-        # use: subshells / arrays
-        if sys.platform == 'win32':  # pragma: win32 cover
-            hook_file.write('#!/bin/sh\n')
+        before = before.replace('#!/usr/bin/env python3', shebang())
 
         hook_file.write(before + TEMPLATE_START)
-        hook_file.write(f'INSTALL_PYTHON={shlex.quote(sys.executable)}\n')
-        args_s = shlex.join(args)
-        hook_file.write(f'ARGS=({args_s})\n')
+        for line in to_template.splitlines():
+            var = line.split()[0]
+            hook_file.write(f'{var} = {params[var]!r}\n')
         hook_file.write(TEMPLATE_END + after)
     make_executable(hook_path)
 
@@ -114,11 +122,11 @@ def _install_hook_script(
 def install(
         config_file: str,
         store: Store,
-        hook_types: list[str] | None,
+        hook_types: Sequence[str],
         overwrite: bool = False,
         hooks: bool = False,
         skip_on_missing_config: bool = False,
-        git_dir: str | None = None,
+        git_dir: Optional[str] = None,
 ) -> int:
     if git_dir is None and git.has_core_hookpaths_set():
         logger.error(
@@ -127,7 +135,7 @@ def install(
         )
         return 1
 
-    for hook_type in _hook_types(config_file, hook_types):
+    for hook_type in hook_types:
         _install_hook_script(
             config_file, hook_type,
             overwrite=overwrite,
@@ -161,7 +169,7 @@ def _uninstall_hook_script(hook_type: str) -> None:
         output.write_line(f'Restored previous hooks to {hook_path}')
 
 
-def uninstall(config_file: str, hook_types: list[str] | None) -> int:
-    for hook_type in _hook_types(config_file, hook_types):
+def uninstall(hook_types: Sequence[str]) -> int:
+    for hook_type in hook_types:
         _uninstall_hook_script(hook_type)
     return 0

pre_commit/commands/migrate_config.py

@@ -1,21 +1,9 @@
-from __future__ import annotations
-
-import functools
-import itertools
+import re
 import textwrap
-from collections.abc import Callable
 
-import cfgv
 import yaml
-from yaml.nodes import ScalarNode
 
-from pre_commit.clientlib import InvalidConfigError
-from pre_commit.yaml import yaml_compose
-from pre_commit.yaml import yaml_load
-from pre_commit.yaml_rewrite import MappingKey
-from pre_commit.yaml_rewrite import MappingValue
-from pre_commit.yaml_rewrite import match
-from pre_commit.yaml_rewrite import SequenceItem
+from pre_commit.util import yaml_load
 
 
 def _is_header_line(line: str) -> bool:
@@ -46,84 +34,16 @@ def _migrate_map(contents: str) -> str:
     return contents
 
 
-def _preserve_style(n: ScalarNode, *, s: str) -> str:
-    style = n.style or ''
-    return f'{style}{s}{style}'
-
-
-def _fix_stage(n: ScalarNode) -> str:
-    return _preserve_style(n, s=f'pre-{n.value}')
-
-
-def _migrate_composed(contents: str) -> str:
-    tree = yaml_compose(contents)
-    rewrites: list[tuple[ScalarNode, Callable[[ScalarNode], str]]] = []
-
-    # sha -> rev
-    sha_to_rev_replace = functools.partial(_preserve_style, s='rev')
-    sha_to_rev_matcher = (
-        MappingValue('repos'),
-        SequenceItem(),
-        MappingKey('sha'),
-    )
-    for node in match(tree, sha_to_rev_matcher):
-        rewrites.append((node, sha_to_rev_replace))
-
-    # python_venv -> python
-    language_matcher = (
-        MappingValue('repos'),
-        SequenceItem(),
-        MappingValue('hooks'),
-        SequenceItem(),
-        MappingValue('language'),
-    )
-    python_venv_replace = functools.partial(_preserve_style, s='python')
-    for node in match(tree, language_matcher):
-        if node.value == 'python_venv':
-            rewrites.append((node, python_venv_replace))
-
-    # stages rewrites
-    default_stages_matcher = (MappingValue('default_stages'), SequenceItem())
-    default_stages_match = match(tree, default_stages_matcher)
-    hook_stages_matcher = (
-        MappingValue('repos'),
-        SequenceItem(),
-        MappingValue('hooks'),
-        SequenceItem(),
-        MappingValue('stages'),
-        SequenceItem(),
-    )
-    hook_stages_match = match(tree, hook_stages_matcher)
-    for node in itertools.chain(default_stages_match, hook_stages_match):
-        if node.value in {'commit', 'push', 'merge-commit'}:
-            rewrites.append((node, _fix_stage))
-
-    rewrites.sort(reverse=True, key=lambda nf: nf[0].start_mark.index)
-
-    src_parts = []
-    end: int | None = None
-    for node, func in rewrites:
-        src_parts.append(contents[node.end_mark.index:end])
-        src_parts.append(func(node))
-        end = node.start_mark.index
-    src_parts.append(contents[:end])
-    src_parts.reverse()
-    return ''.join(src_parts)
+def _migrate_sha_to_rev(contents: str) -> str:
+    return re.sub(r'(\n\s+)sha:', r'\1rev:', contents)
 
 
 def migrate_config(config_file: str, quiet: bool = False) -> int:
     with open(config_file) as f:
         orig_contents = contents = f.read()
 
-    with cfgv.reraise_as(InvalidConfigError):
-        with cfgv.validate_context(f'File {config_file}'):
-            try:
-                yaml_load(orig_contents)
-            except Exception as e:
-                raise cfgv.ValidationError(str(e))
-
     contents = _migrate_map(contents)
-    contents = _migrate_composed(contents)
+    contents = _migrate_sha_to_rev(contents)
 
     if contents != orig_contents:
         with open(config_file, 'w') as f:

pre_commit/commands/run.py

@@ -1,5 +1,3 @@
-from __future__ import annotations
-
 import argparse
 import contextlib
 import functools
@@ -9,20 +7,23 @@ import re
 import subprocess
 import time
 import unicodedata
-from collections.abc import Generator
-from collections.abc import Iterable
-from collections.abc import MutableMapping
-from collections.abc import Sequence
 from typing import Any
+from typing import Collection
+from typing import Dict
+from typing import List
+from typing import MutableMapping
+from typing import Sequence
+from typing import Set
+from typing import Tuple
 
 from identify.identify import tags_from_path
 
 from pre_commit import color
 from pre_commit import git
 from pre_commit import output
-from pre_commit.all_languages import languages
 from pre_commit.clientlib import load_config
 from pre_commit.hook import Hook
+from pre_commit.languages.all import languages
 from pre_commit.repository import all_hooks
 from pre_commit.repository import install_hook_envs
 from pre_commit.staged_files_only import staged_files_only
@@ -58,36 +59,37 @@ def _full_msg(
 
 
 def filter_by_include_exclude(
-        names: Iterable[str],
+        names: Collection[str],
         include: str,
         exclude: str,
-) -> Generator[str]:
+) -> List[str]:
     include_re, exclude_re = re.compile(include), re.compile(exclude)
-    return (
+    return [
         filename for filename in names
         if include_re.search(filename)
         if not exclude_re.search(filename)
-    )
+    ]
 
 
 class Classifier:
-    def __init__(self, filenames: Iterable[str]) -> None:
+    def __init__(self, filenames: Collection[str]) -> None:
         self.filenames = [f for f in filenames if os.path.lexists(f)]
 
-    @functools.cache
-    def _types_for_file(self, filename: str) -> set[str]:
+    @functools.lru_cache(maxsize=None)
+    def _types_for_file(self, filename: str) -> Set[str]:
         return tags_from_path(filename)
 
     def by_types(
             self,
-            names: Iterable[str],
-            types: Iterable[str],
-            types_or: Iterable[str],
-            exclude_types: Iterable[str],
-    ) -> Generator[str]:
+            names: Sequence[str],
+            types: Collection[str],
+            types_or: Collection[str],
+            exclude_types: Collection[str],
+    ) -> List[str]:
         types = frozenset(types)
         types_or = frozenset(types_or)
         exclude_types = frozenset(exclude_types)
+        ret = []
         for filename in names:
             tags = self._types_for_file(filename)
             if (
@@ -95,38 +97,38 @@ class Classifier:
                     (not types_or or tags & types_or) and
                     not tags & exclude_types
             ):
-                yield filename
+                ret.append(filename)
+        return ret
 
-    def filenames_for_hook(self, hook: Hook) -> Generator[str]:
-        return self.by_types(
-            filter_by_include_exclude(
-                self.filenames,
-                hook.files,
-                hook.exclude,
-            ),
+    def filenames_for_hook(self, hook: Hook) -> Tuple[str, ...]:
+        names = self.filenames
+        names = filter_by_include_exclude(names, hook.files, hook.exclude)
+        names = self.by_types(
+            names,
             hook.types,
             hook.types_or,
             hook.exclude_types,
         )
+        return tuple(names)
 
     @classmethod
     def from_config(
             cls,
-            filenames: Iterable[str],
+            filenames: Collection[str],
             include: str,
             exclude: str,
-    ) -> Classifier:
+    ) -> 'Classifier':
         # on windows we normalize all filenames to use forward slashes
         # this makes it easier to filter using the `files:` regex
         # this also makes improperly quoted shell-based hooks work better
         # see #1173
         if os.altsep == '/' and os.sep == '\\':
-            filenames = (f.replace(os.sep, os.altsep) for f in filenames)
+            filenames = [f.replace(os.sep, os.altsep) for f in filenames]
         filenames = filter_by_include_exclude(filenames, include, exclude)
         return Classifier(filenames)
 
 
-def _get_skips(environ: MutableMapping[str, str]) -> set[str]:
+def _get_skips(environ: MutableMapping[str, str]) -> Set[str]:
     skips = environ.get('SKIP', '')
     return {skip.strip() for skip in skips.split(',') if skip.strip()}
 
@@ -142,13 +144,13 @@ def _subtle_line(s: str, use_color: bool) -> None:
 def _run_single_hook(
         classifier: Classifier,
         hook: Hook,
-        skips: set[str],
+        skips: Set[str],
         cols: int,
         diff_before: bytes,
         verbose: bool,
         use_color: bool,
-) -> tuple[bool, bytes]:
-    filenames = tuple(classifier.filenames_for_hook(hook))
+) -> Tuple[bool, bytes]:
+    filenames = classifier.filenames_for_hook(hook)
 
     if hook.id in skips or hook.alias in skips:
         output.write(
@@ -187,19 +189,10 @@ def _run_single_hook(
 
         if not hook.pass_filenames:
             filenames = ()
-        time_before = time.monotonic()
+        time_before = time.time()
         language = languages[hook.language]
-        with language.in_env(hook.prefix, hook.language_version):
-            retcode, out = language.run_hook(
-                hook.prefix,
-                hook.entry,
-                hook.args,
-                filenames,
-                is_local=hook.src == 'local',
-                require_serial=hook.require_serial,
-                color=use_color,
-            )
-        duration = round(time.monotonic() - time_before, 2) or 0
+        retcode, out = language.run_hook(hook, filenames, use_color)
+        duration = round(time.time() - time_before, 2) or 0
         diff_after = _get_diff()
 
         # if the hook makes changes, fail the commit
@@ -250,11 +243,10 @@ def _compute_cols(hooks: Sequence[Hook]) -> int:
     return max(cols, 80)
 
 
-def _all_filenames(args: argparse.Namespace) -> Iterable[str]:
+def _all_filenames(args: argparse.Namespace) -> Collection[str]:
     # these hooks do not operate on files
     if args.hook_stage in {
         'post-checkout', 'post-commit', 'post-merge', 'post-rewrite',
-        'pre-rebase',
     }:
         return ()
     elif args.hook_stage in {'prepare-commit-msg', 'commit-msg'}:
@@ -273,17 +265,17 @@ def _all_filenames(args: argparse.Namespace) -> Iterable[str]:
 
 def _get_diff() -> bytes:
     _, out, _ = cmd_output_b(
-        'git', 'diff', '--no-ext-diff', '--no-textconv', '--ignore-submodules',
-        check=False,
+        'git', 'diff', '--no-ext-diff', '--ignore-submodules', retcode=None,
     )
     return out
 
 
 def _run_hooks(
-        config: dict[str, Any],
+        config: Dict[str, Any],
         hooks: Sequence[Hook],
-        skips: set[str],
+        skips: Set[str],
         args: argparse.Namespace,
+        environ: MutableMapping[str, str],
 ) -> int:
     """Actually run the hooks."""
     cols = _compute_cols(hooks)
@@ -298,8 +290,7 @@ def _run_hooks(
             verbose=args.verbose, use_color=args.color,
         )
         retval |= current_retval
-        fail_fast = (config['fail_fast'] or hook.fail_fast or args.fail_fast)
-        if current_retval and fail_fast:
+        if retval and config['fail_fast']:
             break
     if retval and args.show_diff_on_failure and prior_diff:
         if args.all_files:
@@ -329,7 +320,8 @@ def _has_unmerged_paths() -> bool:
 
 def _has_unstaged_config(config_file: str) -> bool:
     retcode, _, _ = cmd_output_b(
-        'git', 'diff', '--quiet', '--no-ext-diff', config_file, check=False,
+        'git', 'diff', '--no-ext-diff', '--exit-code', config_file,
+        retcode=None,
     )
     # be explicit, other git errors don't mean it has an unstaged config.
     return retcode == 1
@@ -344,7 +336,7 @@ def run(
     stash = not args.all_files and not args.files
 
     # Check if we have unresolved merge conflict files and fail fast.
-    if stash and _has_unmerged_paths():
+    if _has_unmerged_paths():
         logger.error('Unmerged files.  Resolve before committing.')
         return 1
     if bool(args.from_ref) != bool(args.to_ref):
@@ -372,16 +364,6 @@ def run(
     ):
         return 0
 
-    # Expose prepare_commit_message_source / commit_object_name
-    # as environment variables for the hooks
-    if args.prepare_commit_message_source:
-        environ['PRE_COMMIT_COMMIT_MSG_SOURCE'] = (
-            args.prepare_commit_message_source
-        )
-
-    if args.commit_object_name:
-        environ['PRE_COMMIT_COMMIT_OBJECT_NAME'] = args.commit_object_name
-
     # Expose from-ref / to-ref as environment variables for hooks to consume
     if args.from_ref and args.to_ref:
         # legacy names
@@ -391,10 +373,6 @@ def run(
         environ['PRE_COMMIT_FROM_REF'] = args.from_ref
         environ['PRE_COMMIT_TO_REF'] = args.to_ref
 
-    if args.pre_rebase_upstream and args.pre_rebase_branch:
-        environ['PRE_COMMIT_PRE_REBASE_UPSTREAM'] = args.pre_rebase_upstream
-        environ['PRE_COMMIT_PRE_REBASE_BRANCH'] = args.pre_rebase_branch
-
     if (
         args.remote_name and args.remote_url and
         args.remote_branch and args.local_branch
@@ -435,14 +413,10 @@ def run(
             return 1
 
         skips = _get_skips(environ)
-        to_install = [
-            hook
-            for hook in hooks
-            if hook.id not in skips and hook.alias not in skips
-        ]
+        to_install = [hook for hook in hooks if hook.id not in skips]
         install_hook_envs(to_install, store)
 
-        return _run_hooks(config, hooks, skips, args)
+        return _run_hooks(config, hooks, skips, args, environ)
 
     # https://github.com/python/mypy/issues/7726
     raise AssertionError('unreachable')

pre_commit/commands/sample_config.py

@@ -1,4 +1,7 @@
-from __future__ import annotations
+# TODO: maybe `git ls-remote git://github.com/pre-commit/pre-commit-hooks` to
+# determine the latest revision?  This adds ~200ms from my tests (and is
+# significantly faster than https:// or http://).  For now, periodically
+# manually updating the revision is fine.
 SAMPLE_CONFIG = '''\
 # See https://pre-commit.com for more information
 # See https://pre-commit.com/hooks.html for more hooks

pre_commit/commands/try_repo.py

@@ -1,9 +1,8 @@
-from __future__ import annotations
-
 import argparse
 import logging
 import os.path
-import tempfile
+from typing import Optional
+from typing import Tuple
 
 import pre_commit.constants as C
 from pre_commit import git
@@ -12,13 +11,14 @@ from pre_commit.clientlib import load_manifest
 from pre_commit.commands.run import run
 from pre_commit.store import Store
 from pre_commit.util import cmd_output_b
+from pre_commit.util import tmpdir
+from pre_commit.util import yaml_dump
 from pre_commit.xargs import xargs
-from pre_commit.yaml import yaml_dump
 
 logger = logging.getLogger(__name__)
 
 
-def _repo_ref(tmpdir: str, repo: str, ref: str | None) -> tuple[str, str]:
+def _repo_ref(tmpdir: str, repo: str, ref: Optional[str]) -> Tuple[str, str]:
     # if `ref` is explicitly passed, use it
     if ref is not None:
         return repo, ref
@@ -49,7 +49,7 @@ def _repo_ref(tmpdir: str, repo: str, ref: str | None) -> tuple[str, str]:
 
 
 def try_repo(args: argparse.Namespace) -> int:
-    with tempfile.TemporaryDirectory() as tempdir:
+    with tmpdir() as tempdir:
         repo, ref = _repo_ref(tempdir, args.repo, args.ref)
 
         store = Store(tempdir)

pre_commit/commands/validate_config.py

@@ -1,18 +0,0 @@
-from __future__ import annotations
-
-from collections.abc import Sequence
-
-from pre_commit import clientlib
-
-
-def validate_config(filenames: Sequence[str]) -> int:
-    ret = 0
-
-    for filename in filenames:
-        try:
-            clientlib.load_config(filename)
-        except clientlib.InvalidConfigError as e:
-            print(e)
-            ret = 1
-
-    return ret

pre_commit/commands/validate_manifest.py

@@ -1,18 +0,0 @@
-from __future__ import annotations
-
-from collections.abc import Sequence
-
-from pre_commit import clientlib
-
-
-def validate_manifest(filenames: Sequence[str]) -> int:
-    ret = 0
-
-    for filename in filenames:
-        try:
-            clientlib.load_manifest(filename)
-        except clientlib.InvalidManifestError as e:
-            print(e)
-            ret = 1
-
-    return ret

pre_commit/constants.py

@@ -1,13 +1,25 @@
-from __future__ import annotations
+import sys
 
-import importlib.metadata
+if sys.version_info < (3, 8):  # pragma: no cover (<PY38)
+    import importlib_metadata
+else:  # pragma: no cover (PY38+)
+    import importlib.metadata as importlib_metadata
 
 CONFIG_FILE = '.pre-commit-config.yaml'
 MANIFEST_FILE = '.pre-commit-hooks.yaml'
 
+# Bump when installation changes in a backwards / forwards incompatible way
+INSTALLED_STATE_VERSION = '1'
 # Bump when modifying `empty_template`
 LOCAL_REPO_VERSION = '1'
 
-VERSION = importlib.metadata.version('pre_commit')
+VERSION = importlib_metadata.version('pre_commit')
+
+# `manual` is not invoked by any installed git hook.  See #719
+STAGES = (
+    'commit', 'merge-commit', 'prepare-commit-msg', 'commit-msg',
+    'post-commit', 'manual', 'post-checkout', 'push', 'post-merge',
+    'post-rewrite',
+)
 
 DEFAULT = 'default'

pre_commit/envcontext.py

@@ -1,11 +1,11 @@
-from __future__ import annotations
-
 import contextlib
 import enum
 import os
-from collections.abc import Generator
-from collections.abc import MutableMapping
+from typing import Generator
+from typing import MutableMapping
 from typing import NamedTuple
+from typing import Optional
+from typing import Tuple
 from typing import Union
 
 _Unset = enum.Enum('_Unset', 'UNSET')
@@ -17,9 +17,9 @@ class Var(NamedTuple):
     default: str = ''
 
 
-SubstitutionT = tuple[Union[str, Var], ...]
+SubstitutionT = Tuple[Union[str, Var], ...]
 ValueT = Union[str, _Unset, SubstitutionT]
-PatchesT = tuple[tuple[str, ValueT], ...]
+PatchesT = Tuple[Tuple[str, ValueT], ...]
 
 
 def format_env(parts: SubstitutionT, env: MutableMapping[str, str]) -> str:
@@ -32,8 +32,8 @@ def format_env(parts: SubstitutionT, env: MutableMapping[str, str]) -> str:
 @contextlib.contextmanager
 def envcontext(
         patch: PatchesT,
-        _env: MutableMapping[str, str] | None = None,
-) -> Generator[None]:
+        _env: Optional[MutableMapping[str, str]] = None,
+) -> Generator[None, None, None]:
     """In this context, `os.environ` is modified according to `patch`.
 
     `patch` is an iterable of 2-tuples (key, value):

pre_commit/error_handler.py

@@ -1,18 +1,14 @@
-from __future__ import annotations
-
 import contextlib
 import functools
 import os.path
 import sys
 import traceback
-from collections.abc import Generator
-from typing import IO
+from typing import Generator
 
 import pre_commit.constants as C
 from pre_commit import output
 from pre_commit.errors import FatalError
 from pre_commit.store import Store
-from pre_commit.util import cmd_output_b
 from pre_commit.util import force_bytes
 
 
@@ -25,15 +21,12 @@ def _log_and_exit(
     error_msg = f'{msg}: {type(exc).__name__}: '.encode() + force_bytes(exc)
     output.write_line_b(error_msg)
 
-    _, git_version_b, _ = cmd_output_b('git', '--version', check=False)
-    git_version = git_version_b.decode(errors='backslashreplace').rstrip()
-
     storedir = Store().directory
     log_path = os.path.join(storedir, 'pre-commit.log')
     with contextlib.ExitStack() as ctx:
         if os.access(storedir, os.W_OK):
             output.write_line(f'Check the log at {log_path}')
-            log: IO[bytes] = ctx.enter_context(open(log_path, 'wb'))
+            log = ctx.enter_context(open(log_path, 'wb'))
         else:  # pragma: win32 no cover
             output.write_line(f'Failed to write to log at {log_path}')
             log = sys.stdout.buffer
@@ -45,7 +38,6 @@ def _log_and_exit(
         _log_line()
         _log_line('```')
         _log_line(f'pre-commit version: {C.VERSION}')
-        _log_line(f'git --version: {git_version}')
         _log_line('sys.version:')
         for line in sys.version.splitlines():
             _log_line(f'    {line}')
@@ -68,7 +60,7 @@ def _log_and_exit(
 
 
 @contextlib.contextmanager
-def error_handler() -> Generator[None]:
+def error_handler() -> Generator[None, None, None]:
     try:
         yield
     except (Exception, KeyboardInterrupt) as e:

pre_commit/errors.py

@@ -1,5 +1,2 @@
-from __future__ import annotations
-
-
 class FatalError(RuntimeError):
     pass

pre_commit/file_lock.py

@@ -1,10 +1,8 @@
-from __future__ import annotations
-
 import contextlib
 import errno
 import sys
-from collections.abc import Callable
-from collections.abc import Generator
+from typing import Callable
+from typing import Generator
 
 
 if sys.platform == 'win32':  # pragma: no cover (windows)
@@ -20,13 +18,15 @@ if sys.platform == 'win32':  # pragma: no cover (windows)
     def _locked(
             fileno: int,
             blocked_cb: Callable[[], None],
-    ) -> Generator[None]:
+    ) -> Generator[None, None, None]:
         try:
+            # TODO: https://github.com/python/typeshed/pull/3607
             msvcrt.locking(fileno, msvcrt.LK_NBLCK, _region)
         except OSError:
             blocked_cb()
             while True:
                 try:
+                    # TODO: https://github.com/python/typeshed/pull/3607
                     msvcrt.locking(fileno, msvcrt.LK_LOCK, _region)
                 except OSError as e:
                     # Locking violation. Returned when the _LK_LOCK or _LK_RLCK
@@ -45,6 +45,7 @@ if sys.platform == 'win32':  # pragma: no cover (windows)
             # The documentation however states:
             # "Regions should be locked only briefly and should be unlocked
             # before closing a file or exiting the program."
+            # TODO: https://github.com/python/typeshed/pull/3607
             msvcrt.locking(fileno, msvcrt.LK_UNLCK, _region)
 else:  # pragma: win32 no cover
     import fcntl
@@ -53,7 +54,7 @@ else:  # pragma: win32 no cover
     def _locked(
             fileno: int,
             blocked_cb: Callable[[], None],
-    ) -> Generator[None]:
+    ) -> Generator[None, None, None]:
         try:
             fcntl.flock(fileno, fcntl.LOCK_EX | fcntl.LOCK_NB)
         except OSError:  # pragma: no cover (tests are single-threaded)
@@ -69,7 +70,7 @@ else:  # pragma: win32 no cover
 def lock(
         path: str,
         blocked_cb: Callable[[], None],
-) -> Generator[None]:
+) -> Generator[None, None, None]:
     with open(path, 'a+') as f:
         with _locked(f.fileno(), blocked_cb):
             yield

pre_commit/git.py

@@ -1,22 +1,22 @@
-from __future__ import annotations
-
 import logging
 import os.path
 import sys
-from collections.abc import Mapping
+from typing import Dict
+from typing import List
+from typing import MutableMapping
+from typing import Optional
+from typing import Set
 
 from pre_commit.errors import FatalError
 from pre_commit.util import CalledProcessError
 from pre_commit.util import cmd_output
 from pre_commit.util import cmd_output_b
 
+
 logger = logging.getLogger(__name__)
 
-# see #2046
-NO_FS_MONITOR = ('-c', 'core.useBuiltinFSMonitor=false')
 
-
-def zsplit(s: str) -> list[str]:
+def zsplit(s: str) -> List[str]:
     s = s.strip('\0')
     if s:
         return s.split('\0')
@@ -24,7 +24,9 @@ def zsplit(s: str) -> list[str]:
         return []
 
 
-def no_git_env(_env: Mapping[str, str] | None = None) -> dict[str, str]:
+def no_git_env(
+        _env: Optional[MutableMapping[str, str]] = None,
+) -> Dict[str, str]:
     # Too many bugs dealing with environment variables and GIT:
     # https://github.com/pre-commit/pre-commit/issues/300
     # In git 2.6.3 (maybe others), git exports GIT_WORK_TREE while running
@@ -37,13 +39,9 @@ def no_git_env(_env: Mapping[str, str] | None = None) -> dict[str, str]:
     return {
         k: v for k, v in _env.items()
         if not k.startswith('GIT_') or
-        k.startswith(('GIT_CONFIG_KEY_', 'GIT_CONFIG_VALUE_')) or
         k in {
             'GIT_EXEC_PATH', 'GIT_SSH', 'GIT_SSH_COMMAND', 'GIT_SSL_CAINFO',
-            'GIT_SSL_NO_VERIFY', 'GIT_CONFIG_COUNT',
-            'GIT_HTTP_PROXY_AUTHMETHOD',
-            'GIT_ALLOW_PROTOCOL',
-            'GIT_ASKPASS',
+            'GIT_SSL_NO_VERIFY',
         }
     }
 
@@ -57,15 +55,13 @@ def get_root() -> str:
         root = os.path.abspath(
             cmd_output('git', 'rev-parse', '--show-cdup')[1].strip(),
         )
-        inside_git_dir = cmd_output(
-            'git', 'rev-parse', '--is-inside-git-dir',
-        )[1].strip()
+        git_dir = os.path.abspath(get_git_dir())
     except CalledProcessError:
         raise FatalError(
             'git failed. Is it installed, and are you in a Git repository '
             'directory?',
         )
-    if inside_git_dir != 'false':
+    if os.path.samefile(root, git_dir):
         raise FatalError(
             'git toplevel unexpectedly empty! make sure you are not '
             'inside the `.git` directory of your repository.',
@@ -74,23 +70,18 @@ def get_root() -> str:
 
 
 def get_git_dir(git_root: str = '.') -> str:
-    opt = '--git-dir'
-    _, out, _ = cmd_output('git', 'rev-parse', opt, cwd=git_root)
-    git_dir = out.strip()
-    if git_dir != opt:
-        return os.path.normpath(os.path.join(git_root, git_dir))
+    opts = ('--git-common-dir', '--git-dir')
+    _, out, _ = cmd_output('git', 'rev-parse', *opts, cwd=git_root)
+    for line, opt in zip(out.splitlines(), opts):
+        if line != opt:  # pragma: no branch (git < 2.5)
+            return os.path.normpath(os.path.join(git_root, line))
     else:
         raise AssertionError('unreachable: no git dir')
 
 
-def get_git_common_dir(git_root: str = '.') -> str:
-    opt = '--git-common-dir'
-    _, out, _ = cmd_output('git', 'rev-parse', opt, cwd=git_root)
-    git_common_dir = out.strip()
-    if git_common_dir != opt:
-        return os.path.normpath(os.path.join(git_root, git_common_dir))
-    else:  # pragma: no cover (git < 2.5)
-        return get_git_dir(git_root)
+def get_remote_url(git_root: str) -> str:
+    _, out, _ = cmd_output('git', 'config', 'remote.origin.url', cwd=git_root)
+    return out.strip()
 
 
 def is_in_merge_conflict() -> bool:
@@ -101,7 +92,7 @@ def is_in_merge_conflict() -> bool:
     )
 
 
-def parse_merge_msg_for_conflicts(merge_msg: bytes) -> list[str]:
+def parse_merge_msg_for_conflicts(merge_msg: bytes) -> List[str]:
     # Conflicted files start with tabs
     return [
         line.lstrip(b'#').strip().decode()
@@ -111,7 +102,7 @@ def parse_merge_msg_for_conflicts(merge_msg: bytes) -> list[str]:
     ]
 
 
-def get_conflicted_files() -> set[str]:
+def get_conflicted_files() -> Set[str]:
     logger.info('Checking merge-conflict files only.')
     # Need to get the conflicted files from the MERGE_MSG because they could
     # have resolved the conflict by choosing one side or the other
@@ -126,13 +117,13 @@ def get_conflicted_files() -> set[str]:
     merge_diff_filenames = zsplit(
         cmd_output(
             'git', 'diff', '--name-only', '--no-ext-diff', '-z',
-            '-m', tree_hash, 'HEAD', 'MERGE_HEAD', '--',
+            '-m', tree_hash, 'HEAD', 'MERGE_HEAD',
         )[1],
     )
     return set(merge_conflict_filenames) | set(merge_diff_filenames)
 
 
-def get_staged_files(cwd: str | None = None) -> list[str]:
+def get_staged_files(cwd: Optional[str] = None) -> List[str]:
     return zsplit(
         cmd_output(
             'git', 'diff', '--staged', '--name-only', '--no-ext-diff', '-z',
@@ -143,19 +134,27 @@ def get_staged_files(cwd: str | None = None) -> list[str]:
     )
 
 
-def intent_to_add_files() -> list[str]:
+def intent_to_add_files() -> List[str]:
     _, stdout, _ = cmd_output(
-        'git', 'diff', '--no-ext-diff', '--ignore-submodules',
-        '--diff-filter=A', '--name-only', '-z',
+        'git', 'status', '--ignore-submodules', '--porcelain', '-z',
     )
-    return zsplit(stdout)
+    parts = list(reversed(zsplit(stdout)))
+    intent_to_add = []
+    while parts:
+        line = parts.pop()
+        status, filename = line[:3], line[3:]
+        if status[0] in {'C', 'R'}:  # renames / moves have an additional arg
+            parts.pop()
+        if status[1] == 'A':
+            intent_to_add.append(filename)
+    return intent_to_add
 
 
-def get_all_files() -> list[str]:
+def get_all_files() -> List[str]:
     return zsplit(cmd_output('git', 'ls-files', '-z')[1])
 
 
-def get_changed_files(old: str, new: str) -> list[str]:
+def get_changed_files(old: str, new: str) -> List[str]:
     diff_cmd = ('git', 'diff', '--name-only', '--no-ext-diff', '-z')
     try:
         _, out, _ = cmd_output(*diff_cmd, f'{old}...{new}')
@@ -174,11 +173,11 @@ def head_rev(remote: str) -> str:
 
 def has_diff(*args: str, repo: str = '.') -> bool:
     cmd = ('git', 'diff', '--quiet', '--no-ext-diff', *args)
-    return cmd_output_b(*cmd, cwd=repo, check=False)[0] == 1
+    return cmd_output_b(*cmd, cwd=repo, retcode=None)[0] == 1
 
 
 def has_core_hookpaths_set() -> bool:
-    _, out, _ = cmd_output_b('git', 'config', 'core.hooksPath', check=False)
+    _, out, _ = cmd_output_b('git', 'config', 'core.hooksPath', retcode=None)
     return bool(out.strip())
 
 
@@ -186,11 +185,10 @@ def init_repo(path: str, remote: str) -> None:
     if os.path.isdir(remote):
         remote = os.path.abspath(remote)
 
-    git = ('git', *NO_FS_MONITOR)
     env = no_git_env()
     # avoid the user's template so that hooks do not recurse
-    cmd_output_b(*git, 'init', '--template=', path, env=env)
-    cmd_output_b(*git, 'remote', 'add', 'origin', remote, cwd=path, env=env)
+    cmd_output_b('git', 'init', '--template=', path, env=env)
+    cmd_output_b('git', 'remote', 'add', 'origin', remote, cwd=path, env=env)
 
 
 def commit(repo: str = '.') -> None:
@@ -219,7 +217,7 @@ def check_for_cygwin_mismatch() -> None:
 
         if is_cygwin_python ^ is_cygwin_git:
             exe_type = {True: '(cygwin)', False: '(windows)'}
-            logger.warning(
+            logger.warn(
                 f'pre-commit has detected a mix of cygwin python / git\n'
                 f'This combination is not supported, it is likely you will '
                 f'receive an error later in the program.\n'
@@ -228,18 +226,3 @@ def check_for_cygwin_mismatch() -> None:
                 f' - python {exe_type[is_cygwin_python]}\n'
                 f' - git {exe_type[is_cygwin_git]}\n',
             )
-
-
-def get_best_candidate_tag(rev: str, git_repo: str) -> str:
-    """Get the best tag candidate.
-
-    Multiple tags can exist on a SHA. Sometimes a moving tag is attached
-    to a version tag. Try to pick the tag that looks like a version.
-    """
-    tags = cmd_output(
-        'git', *NO_FS_MONITOR, 'tag', '--points-at', rev, cwd=git_repo,
-    )[1].splitlines()
-    for tag in tags:
-        if '.' in tag:
-            return tag
-    return rev

pre_commit/hook.py

@@ -1,9 +1,10 @@
-from __future__ import annotations
-
 import logging
-from collections.abc import Sequence
+import shlex
 from typing import Any
+from typing import Dict
 from typing import NamedTuple
+from typing import Sequence
+from typing import Tuple
 
 from pre_commit.prefix import Prefix
 
@@ -26,7 +27,6 @@ class Hook(NamedTuple):
     additional_dependencies: Sequence[str]
     args: Sequence[str]
     always_run: bool
-    fail_fast: bool
     pass_filenames: bool
     description: str
     language_version: str
@@ -37,7 +37,11 @@ class Hook(NamedTuple):
     verbose: bool
 
     @property
-    def install_key(self) -> tuple[Prefix, str, str, tuple[str, ...]]:
+    def cmd(self) -> Tuple[str, ...]:
+        return (*shlex.split(self.entry), *self.args)
+
+    @property
+    def install_key(self) -> Tuple[Prefix, str, str, Tuple[str, ...]]:
         return (
             self.prefix,
             self.language,
@@ -46,7 +50,7 @@ class Hook(NamedTuple):
         )
 
     @classmethod
-    def create(cls, src: str, prefix: Prefix, dct: dict[str, Any]) -> Hook:
+    def create(cls, src: str, prefix: Prefix, dct: Dict[str, Any]) -> 'Hook':
         # TODO: have cfgv do this (?)
         extra_keys = set(dct) - _KEYS
         if extra_keys:

pre_commit/lang_base.py

@@ -1,196 +0,0 @@
-from __future__ import annotations
-
-import contextlib
-import os
-import random
-import re
-import shlex
-import sys
-from collections.abc import Generator
-from collections.abc import Sequence
-from typing import Any
-from typing import ContextManager
-from typing import NoReturn
-from typing import Protocol
-
-import pre_commit.constants as C
-from pre_commit import parse_shebang
-from pre_commit import xargs
-from pre_commit.prefix import Prefix
-from pre_commit.util import cmd_output_b
-
-FIXED_RANDOM_SEED = 1542676187
-
-SHIMS_RE = re.compile(r'[/\\]shims[/\\]')
-
-
-class Language(Protocol):
-    # Use `None` for no installation / environment
-    @property
-    def ENVIRONMENT_DIR(self) -> str | None: ...
-    # return a value to replace `'default` for `language_version`
-    def get_default_version(self) -> str: ...
-    # return whether the environment is healthy (or should be rebuilt)
-    def health_check(self, prefix: Prefix, version: str) -> str | None: ...
-
-    # install a repository for the given language and language_version
-    def install_environment(
-            self,
-            prefix: Prefix,
-            version: str,
-            additional_dependencies: Sequence[str],
-    ) -> None:
-        ...
-
-    # modify the environment for hook execution
-    def in_env(self, prefix: Prefix, version: str) -> ContextManager[None]: ...
-
-    # execute a hook and return the exit code and output
-    def run_hook(
-            self,
-            prefix: Prefix,
-            entry: str,
-            args: Sequence[str],
-            file_args: Sequence[str],
-            *,
-            is_local: bool,
-            require_serial: bool,
-            color: bool,
-    ) -> tuple[int, bytes]:
-        ...
-
-
-def exe_exists(exe: str) -> bool:
-    found = parse_shebang.find_executable(exe)
-    if found is None:  # exe exists
-        return False
-
-    homedir = os.path.expanduser('~')
-    try:
-        common: str | None = os.path.commonpath((found, homedir))
-    except ValueError:  # on windows, different drives raises ValueError
-        common = None
-
-    return (
-        # it is not in a /shims/ directory
-        not SHIMS_RE.search(found) and
-        (
-            # the homedir is / (docker, service user, etc.)
-            os.path.dirname(homedir) == homedir or
-            # the exe is not contained in the home directory
-            common != homedir
-        )
-    )
-
-
-def setup_cmd(prefix: Prefix, cmd: tuple[str, ...], **kwargs: Any) -> None:
-    cmd_output_b(*cmd, cwd=prefix.prefix_dir, **kwargs)
-
-
-def environment_dir(prefix: Prefix, d: str, language_version: str) -> str:
-    return prefix.path(f'{d}-{language_version}')
-
-
-def assert_version_default(binary: str, version: str) -> None:
-    if version != C.DEFAULT:
-        raise AssertionError(
-            f'for now, pre-commit requires system-installed {binary} -- '
-            f'you selected `language_version: {version}`',
-        )
-
-
-def assert_no_additional_deps(
-        lang: str,
-        additional_deps: Sequence[str],
-) -> None:
-    if additional_deps:
-        raise AssertionError(
-            f'for now, pre-commit does not support '
-            f'additional_dependencies for {lang} -- '
-            f'you selected `additional_dependencies: {additional_deps}`',
-        )
-
-
-def basic_get_default_version() -> str:
-    return C.DEFAULT
-
-
-def basic_health_check(prefix: Prefix, language_version: str) -> str | None:
-    return None
-
-
-def no_install(
-        prefix: Prefix,
-        version: str,
-        additional_dependencies: Sequence[str],
-) -> NoReturn:
-    raise AssertionError('This language is not installable')
-
-
-@contextlib.contextmanager
-def no_env(prefix: Prefix, version: str) -> Generator[None]:
-    yield
-
-
-def target_concurrency() -> int:
-    if 'PRE_COMMIT_NO_CONCURRENCY' in os.environ:
-        return 1
-    else:
-        # Travis appears to have a bunch of CPUs, but we can't use them all.
-        if 'TRAVIS' in os.environ:
-            return 2
-        else:
-            return xargs.cpu_count()
-
-
-def _shuffled(seq: Sequence[str]) -> list[str]:
-    """Deterministically shuffle"""
-    fixed_random = random.Random()
-    fixed_random.seed(FIXED_RANDOM_SEED, version=1)
-
-    seq = list(seq)
-    fixed_random.shuffle(seq)
-    return seq
-
-
-def run_xargs(
-        cmd: tuple[str, ...],
-        file_args: Sequence[str],
-        *,
-        require_serial: bool,
-        color: bool,
-) -> tuple[int, bytes]:
-    if require_serial:
-        jobs = 1
-    else:
-        # Shuffle the files so that they more evenly fill out the xargs
-        # partitions, but do it deterministically in case a hook cares about
-        # ordering.
-        file_args = _shuffled(file_args)
-        jobs = target_concurrency()
-    return xargs.xargs(cmd, file_args, target_concurrency=jobs, color=color)
-
-
-def hook_cmd(entry: str, args: Sequence[str]) -> tuple[str, ...]:
-    cmd = shlex.split(entry)
-    if cmd[:2] == ['pre-commit', 'hazmat']:
-        cmd = [sys.executable, '-m', 'pre_commit.commands.hazmat', *cmd[2:]]
-    return (*cmd, *args)
-
-
-def basic_run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
-        file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
-        color: bool,
-) -> tuple[int, bytes]:
-    return run_xargs(
-        hook_cmd(entry, args),
-        file_args,
-        require_serial=require_serial,
-        color=color,
-    )

pre_commit/languages/all.py

@@ -0,0 +1,68 @@
+from typing import Callable
+from typing import NamedTuple
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
+
+from pre_commit.hook import Hook
+from pre_commit.languages import conda
+from pre_commit.languages import coursier
+from pre_commit.languages import dart
+from pre_commit.languages import docker
+from pre_commit.languages import docker_image
+from pre_commit.languages import dotnet
+from pre_commit.languages import fail
+from pre_commit.languages import golang
+from pre_commit.languages import node
+from pre_commit.languages import perl
+from pre_commit.languages import pygrep
+from pre_commit.languages import python
+from pre_commit.languages import r
+from pre_commit.languages import ruby
+from pre_commit.languages import rust
+from pre_commit.languages import script
+from pre_commit.languages import swift
+from pre_commit.languages import system
+from pre_commit.prefix import Prefix
+
+
+class Language(NamedTuple):
+    name: str
+    # Use `None` for no installation / environment
+    ENVIRONMENT_DIR: Optional[str]
+    # return a value to replace `'default` for `language_version`
+    get_default_version: Callable[[], str]
+    # return whether the environment is healthy (or should be rebuilt)
+    healthy: Callable[[Prefix, str], bool]
+    # install a repository for the given language and language_version
+    install_environment: Callable[[Prefix, str, Sequence[str]], None]
+    # execute a hook and return the exit code and output
+    run_hook: 'Callable[[Hook, Sequence[str], bool], Tuple[int, bytes]]'
+
+
+# TODO: back to modules + Protocol: https://github.com/python/mypy/issues/5018
+languages = {
+    # BEGIN GENERATED (testing/gen-languages-all)
+    'conda': Language(name='conda', ENVIRONMENT_DIR=conda.ENVIRONMENT_DIR, get_default_version=conda.get_default_version, healthy=conda.healthy, install_environment=conda.install_environment, run_hook=conda.run_hook),  # noqa: E501
+    'coursier': Language(name='coursier', ENVIRONMENT_DIR=coursier.ENVIRONMENT_DIR, get_default_version=coursier.get_default_version, healthy=coursier.healthy, install_environment=coursier.install_environment, run_hook=coursier.run_hook),  # noqa: E501
+    'dart': Language(name='dart', ENVIRONMENT_DIR=dart.ENVIRONMENT_DIR, get_default_version=dart.get_default_version, healthy=dart.healthy, install_environment=dart.install_environment, run_hook=dart.run_hook),  # noqa: E501
+    'docker': Language(name='docker', ENVIRONMENT_DIR=docker.ENVIRONMENT_DIR, get_default_version=docker.get_default_version, healthy=docker.healthy, install_environment=docker.install_environment, run_hook=docker.run_hook),  # noqa: E501
+    'docker_image': Language(name='docker_image', ENVIRONMENT_DIR=docker_image.ENVIRONMENT_DIR, get_default_version=docker_image.get_default_version, healthy=docker_image.healthy, install_environment=docker_image.install_environment, run_hook=docker_image.run_hook),  # noqa: E501
+    'dotnet': Language(name='dotnet', ENVIRONMENT_DIR=dotnet.ENVIRONMENT_DIR, get_default_version=dotnet.get_default_version, healthy=dotnet.healthy, install_environment=dotnet.install_environment, run_hook=dotnet.run_hook),  # noqa: E501
+    'fail': Language(name='fail', ENVIRONMENT_DIR=fail.ENVIRONMENT_DIR, get_default_version=fail.get_default_version, healthy=fail.healthy, install_environment=fail.install_environment, run_hook=fail.run_hook),  # noqa: E501
+    'golang': Language(name='golang', ENVIRONMENT_DIR=golang.ENVIRONMENT_DIR, get_default_version=golang.get_default_version, healthy=golang.healthy, install_environment=golang.install_environment, run_hook=golang.run_hook),  # noqa: E501
+    'node': Language(name='node', ENVIRONMENT_DIR=node.ENVIRONMENT_DIR, get_default_version=node.get_default_version, healthy=node.healthy, install_environment=node.install_environment, run_hook=node.run_hook),  # noqa: E501
+    'perl': Language(name='perl', ENVIRONMENT_DIR=perl.ENVIRONMENT_DIR, get_default_version=perl.get_default_version, healthy=perl.healthy, install_environment=perl.install_environment, run_hook=perl.run_hook),  # noqa: E501
+    'pygrep': Language(name='pygrep', ENVIRONMENT_DIR=pygrep.ENVIRONMENT_DIR, get_default_version=pygrep.get_default_version, healthy=pygrep.healthy, install_environment=pygrep.install_environment, run_hook=pygrep.run_hook),  # noqa: E501
+    'python': Language(name='python', ENVIRONMENT_DIR=python.ENVIRONMENT_DIR, get_default_version=python.get_default_version, healthy=python.healthy, install_environment=python.install_environment, run_hook=python.run_hook),  # noqa: E501
+    'r': Language(name='r', ENVIRONMENT_DIR=r.ENVIRONMENT_DIR, get_default_version=r.get_default_version, healthy=r.healthy, install_environment=r.install_environment, run_hook=r.run_hook),  # noqa: E501
+    'ruby': Language(name='ruby', ENVIRONMENT_DIR=ruby.ENVIRONMENT_DIR, get_default_version=ruby.get_default_version, healthy=ruby.healthy, install_environment=ruby.install_environment, run_hook=ruby.run_hook),  # noqa: E501
+    'rust': Language(name='rust', ENVIRONMENT_DIR=rust.ENVIRONMENT_DIR, get_default_version=rust.get_default_version, healthy=rust.healthy, install_environment=rust.install_environment, run_hook=rust.run_hook),  # noqa: E501
+    'script': Language(name='script', ENVIRONMENT_DIR=script.ENVIRONMENT_DIR, get_default_version=script.get_default_version, healthy=script.healthy, install_environment=script.install_environment, run_hook=script.run_hook),  # noqa: E501
+    'swift': Language(name='swift', ENVIRONMENT_DIR=swift.ENVIRONMENT_DIR, get_default_version=swift.get_default_version, healthy=swift.healthy, install_environment=swift.install_environment, run_hook=swift.run_hook),  # noqa: E501
+    'system': Language(name='system', ENVIRONMENT_DIR=system.ENVIRONMENT_DIR, get_default_version=system.get_default_version, healthy=system.healthy, install_environment=system.install_environment, run_hook=system.run_hook),  # noqa: E501
+    # END GENERATED
+}
+# TODO: fully deprecate `python_venv`
+languages['python_venv'] = languages['python']
+all_languages = sorted(languages)

pre_commit/languages/conda.py

@@ -1,24 +1,23 @@
-from __future__ import annotations
-
 import contextlib
 import os
-import sys
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import SubstitutionT
 from pre_commit.envcontext import UNSET
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output_b
 
 ENVIRONMENT_DIR = 'conda'
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
 def get_env_patch(env: str) -> PatchesT:
@@ -27,7 +26,7 @@ def get_env_patch(env: str) -> PatchesT:
     # $CONDA_PREFIX/Scripts and $CONDA_PREFIX. Whereas the latter only
     # seems to be used for python.exe.
     path: SubstitutionT = (os.path.join(env, 'bin'), os.pathsep, Var('PATH'))
-    if sys.platform == 'win32':  # pragma: win32 cover
+    if os.name == 'nt':  # pragma: no cover (platform specific)
         path = (env, os.pathsep, *path)
         path = (os.path.join(env, 'Scripts'), os.pathsep, *path)
         path = (os.path.join(env, 'Library', 'bin'), os.pathsep, *path)
@@ -41,37 +40,45 @@ def get_env_patch(env: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+def in_env(
+        prefix: Prefix,
+        language_version: str,
+) -> Generator[None, None, None]:
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, language_version)
+    envdir = prefix.path(directory)
     with envcontext(get_env_patch(envdir)):
         yield
 
 
-def _conda_exe() -> str:
-    if os.environ.get('PRE_COMMIT_USE_MICROMAMBA'):
-        return 'micromamba'
-    elif os.environ.get('PRE_COMMIT_USE_MAMBA'):
-        return 'mamba'
-    else:
-        return 'conda'
-
-
 def install_environment(
         prefix: Prefix,
         version: str,
         additional_dependencies: Sequence[str],
 ) -> None:
-    lang_base.assert_version_default('conda', version)
+    helpers.assert_version_default('conda', version)
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, version)
 
-    conda_exe = _conda_exe()
-
-    env_dir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    cmd_output_b(
-        conda_exe, 'env', 'create', '-p', env_dir, '--file',
-        'environment.yml', cwd=prefix.prefix_dir,
-    )
-    if additional_dependencies:
+    env_dir = prefix.path(directory)
+    with clean_path_on_failure(env_dir):
         cmd_output_b(
-            conda_exe, 'install', '-p', env_dir, *additional_dependencies,
-            cwd=prefix.prefix_dir,
+            'conda', 'env', 'create', '-p', env_dir, '--file',
+            'environment.yml', cwd=prefix.prefix_dir,
         )
+        if additional_dependencies:
+            cmd_output_b(
+                'conda', 'install', '-p', env_dir, *additional_dependencies,
+                cwd=prefix.prefix_dir,
+            )
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    # TODO: Some rare commands need to be run using `conda run` but mostly we
+    #       can run them without which is much quicker and produces a better
+    #       output.
+    # cmd = ('conda', 'run', '-p', env_dir) + hook.cmd
+    with in_env(hook.prefix, hook.language_version):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/coursier.py

@@ -1,76 +1,71 @@
-from __future__ import annotations
-
 import contextlib
-import os.path
-from collections.abc import Generator
-from collections.abc import Sequence
+import os
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import Var
-from pre_commit.errors import FatalError
-from pre_commit.parse_shebang import find_executable
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 
 ENVIRONMENT_DIR = 'coursier'
 
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
 def install_environment(
         prefix: Prefix,
         version: str,
         additional_dependencies: Sequence[str],
-) -> None:
-    lang_base.assert_version_default('coursier', version)
+) -> None:   # pragma: win32 no cover
+    helpers.assert_version_default('coursier', version)
+    helpers.assert_no_additional_deps('coursier', additional_dependencies)
 
-    # Support both possible executable names (either "cs" or "coursier")
-    cs = find_executable('cs') or find_executable('coursier')
-    if cs is None:
-        raise AssertionError(
-            'pre-commit requires system-installed "cs" or "coursier" '
-            'executables in the application search path',
-        )
-
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-
-    def _install(*opts: str) -> None:
-        assert cs is not None
-        lang_base.setup_cmd(prefix, (cs, 'fetch', *opts))
-        lang_base.setup_cmd(prefix, (cs, 'install', '--dir', envdir, *opts))
-
-    with in_env(prefix, version):
-        channel = prefix.path('.pre-commit-channel')
-        if os.path.isdir(channel):
-            for app_descriptor in os.listdir(channel):
-                _, app_file = os.path.split(app_descriptor)
-                app, _ = os.path.splitext(app_file)
-                _install(
+    envdir = prefix.path(helpers.environment_dir(ENVIRONMENT_DIR, version))
+    channel = prefix.path('.pre-commit-channel')
+    with clean_path_on_failure(envdir):
+        for app_descriptor in os.listdir(channel):
+            _, app_file = os.path.split(app_descriptor)
+            app, _ = os.path.splitext(app_file)
+            helpers.run_setup_cmd(
+                prefix,
+                (
+                    'cs',
+                    'install',
                     '--default-channels=false',
-                    '--channel', channel,
+                    f'--channel={channel}',
                     app,
-                )
-        elif not additional_dependencies:
-            raise FatalError(
-                'expected .pre-commit-channel dir or additional_dependencies',
+                    f'--dir={envdir}',
+                ),
             )
 
-        if additional_dependencies:
-            _install(*additional_dependencies)
 
-
-def get_env_patch(target_dir: str) -> PatchesT:
+def get_env_patch(target_dir: str) -> PatchesT:   # pragma: win32 no cover
     return (
         ('PATH', (target_dir, os.pathsep, Var('PATH'))),
-        ('COURSIER_CACHE', os.path.join(target_dir, '.cs-cache')),
     )
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir)):
+def in_env(
+        prefix: Prefix,
+) -> Generator[None, None, None]:   # pragma: win32 no cover
+    target_dir = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, get_default_version()),
+    )
+    with envcontext(get_env_patch(target_dir)):
         yield
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:   # pragma: win32 no cover
+    with in_env(hook.prefix):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/dart.py

@@ -1,25 +1,26 @@
-from __future__ import annotations
-
 import contextlib
 import os.path
 import shutil
 import tempfile
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
+import pre_commit.constants as C
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import win_exe
-from pre_commit.yaml import yaml_load
+from pre_commit.util import yaml_load
 
 ENVIRONMENT_DIR = 'dartenv'
 
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
 def get_env_patch(venv: str) -> PatchesT:
@@ -29,8 +30,9 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+def in_env(prefix: Prefix) -> Generator[None, None, None]:
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT)
+    envdir = prefix.path(directory)
     with envcontext(get_env_patch(envdir)):
         yield
 
@@ -40,9 +42,9 @@ def install_environment(
         version: str,
         additional_dependencies: Sequence[str],
 ) -> None:
-    lang_base.assert_version_default('dart', version)
+    helpers.assert_version_default('dart', version)
 
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    envdir = prefix.path(helpers.environment_dir(ENVIRONMENT_DIR, version))
     bin_dir = os.path.join(envdir, 'bin')
 
     def _install_dir(prefix_p: Prefix, pub_cache: str) -> None:
@@ -51,10 +53,10 @@ def install_environment(
         with open(prefix_p.path('pubspec.yaml')) as f:
             pubspec_contents = yaml_load(f)
 
-        lang_base.setup_cmd(prefix_p, ('dart', 'pub', 'get'), env=dart_env)
+        helpers.run_setup_cmd(prefix_p, ('dart', 'pub', 'get'), env=dart_env)
 
         for executable in pubspec_contents['executables']:
-            lang_base.setup_cmd(
+            helpers.run_setup_cmd(
                 prefix_p,
                 (
                     'dart', 'compile', 'exe',
@@ -64,34 +66,44 @@ def install_environment(
                 env=dart_env,
             )
 
-    os.makedirs(bin_dir)
+    with clean_path_on_failure(envdir):
+        os.makedirs(bin_dir)
 
-    with tempfile.TemporaryDirectory() as tmp:
-        _install_dir(prefix, tmp)
+        with tempfile.TemporaryDirectory() as tmp:
+            _install_dir(prefix, tmp)
 
-    for dep_s in additional_dependencies:
-        with tempfile.TemporaryDirectory() as dep_tmp:
-            dep, _, version = dep_s.partition(':')
-            if version:
-                dep_cmd: tuple[str, ...] = (dep, '--version', version)
-            else:
-                dep_cmd = (dep,)
+        for dep_s in additional_dependencies:
+            with tempfile.TemporaryDirectory() as dep_tmp:
+                dep, _, version = dep_s.partition(':')
+                if version:
+                    dep_cmd: Tuple[str, ...] = (dep, '--version', version)
+                else:
+                    dep_cmd = (dep,)
 
-            lang_base.setup_cmd(
-                prefix,
-                ('dart', 'pub', 'cache', 'add', *dep_cmd),
-                env={**os.environ, 'PUB_CACHE': dep_tmp},
-            )
-
-            # try and find the 'pubspec.yaml' that just got added
-            for root, _, filenames in os.walk(dep_tmp):
-                if 'pubspec.yaml' in filenames:
-                    with tempfile.TemporaryDirectory() as copied:
-                        pkg = os.path.join(copied, 'pkg')
-                        shutil.copytree(root, pkg)
-                        _install_dir(Prefix(pkg), dep_tmp)
-                    break
-            else:
-                raise AssertionError(
-                    f'could not find pubspec.yaml for {dep_s}',
+                helpers.run_setup_cmd(
+                    prefix,
+                    ('dart', 'pub', 'cache', 'add', *dep_cmd),
+                    env={**os.environ, 'PUB_CACHE': dep_tmp},
                 )
+
+                # try and find the 'pubspec.yaml' that just got added
+                for root, _, filenames in os.walk(dep_tmp):
+                    if 'pubspec.yaml' in filenames:
+                        with tempfile.TemporaryDirectory() as copied:
+                            pkg = os.path.join(copied, 'pkg')
+                            shutil.copytree(root, pkg)
+                            _install_dir(Prefix(pkg), dep_tmp)
+                        break
+                else:
+                    raise AssertionError(
+                        f'could not find pubspec.yaml for {dep_s}',
+                    )
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/docker.py

@@ -1,52 +1,48 @@
-from __future__ import annotations
-
-import contextlib
-import functools
 import hashlib
 import json
 import os
-import re
-from collections.abc import Sequence
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
+import pre_commit.constants as C
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
 from pre_commit.util import CalledProcessError
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output_b
 
 ENVIRONMENT_DIR = 'docker'
 PRE_COMMIT_LABEL = 'PRE_COMMIT'
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-in_env = lang_base.no_env  # no special environment for docker
-
-_HOSTNAME_MOUNT_RE = re.compile(
-    rb"""
-    /containers
-    (?:/overlay-containers)?
-    /([a-z0-9]{64})
-    (?:/userdata)?
-    /hostname
-    """,
-    re.VERBOSE,
-)
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
-def _get_container_id() -> str | None:
-    with contextlib.suppress(FileNotFoundError):
-        with open('/proc/1/mountinfo', 'rb') as f:
-            for line in f:
-                m = _HOSTNAME_MOUNT_RE.search(line)
-                if m:
-                    return m[1].decode()
+def _is_in_docker() -> bool:
+    try:
+        with open('/proc/1/cgroup', 'rb') as f:
+            return b'docker' in f.read()
+    except FileNotFoundError:
+        return False
 
-    return None
+
+def _get_container_id() -> str:
+    # It's assumed that we already check /proc/1/cgroup in _is_in_docker. The
+    # cpuset cgroup controller existed since cgroups were introduced so this
+    # way of getting the container ID is pretty reliable.
+    with open('/proc/1/cgroup', 'rb') as f:
+        for line in f.readlines():
+            if line.split(b':')[1] == b'cpuset':
+                return os.path.basename(line.split(b':')[2]).strip().decode()
+    raise RuntimeError('Failed to find the container ID in /proc/1/cgroup.')
 
 
 def _get_docker_path(path: str) -> str:
-    container_id = _get_container_id()
-    if container_id is None:
+    if not _is_in_docker():
         return path
 
+    container_id = _get_container_id()
+
     try:
         _, out, _ = cmd_output_b('docker', 'inspect', container_id)
     except CalledProcessError:
@@ -80,7 +76,7 @@ def build_docker_image(
         *,
         pull: bool,
 ) -> None:  # pragma: win32 no cover
-    cmd: tuple[str, ...] = (
+    cmd: Tuple[str, ...] = (
         'docker', 'build',
         '--tag', docker_tag(prefix),
         '--label', PRE_COMMIT_LABEL,
@@ -89,64 +85,37 @@ def build_docker_image(
         cmd += ('--pull',)
     # This must come last for old versions of docker.  See #477
     cmd += ('.',)
-    lang_base.setup_cmd(prefix, cmd)
+    helpers.run_setup_cmd(prefix, cmd)
 
 
 def install_environment(
         prefix: Prefix, version: str, additional_dependencies: Sequence[str],
 ) -> None:  # pragma: win32 no cover
-    lang_base.assert_version_default('docker', version)
-    lang_base.assert_no_additional_deps('docker', additional_dependencies)
+    helpers.assert_version_default('docker', version)
+    helpers.assert_no_additional_deps('docker', additional_dependencies)
 
-    directory = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    directory = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT),
+    )
 
     # Docker doesn't really have relevant disk environment, but pre-commit
     # still needs to cleanup its state files on failure
-    build_docker_image(prefix, pull=True)
-    os.mkdir(directory)
+    with clean_path_on_failure(directory):
+        build_docker_image(prefix, pull=True)
+        os.mkdir(directory)
 
 
-@functools.lru_cache(maxsize=1)
-def _is_rootless() -> bool:  # pragma: win32 no cover
-    retcode, out, _ = cmd_output_b(
-        'docker', 'system', 'info', '--format', '{{ json . }}',
-    )
-    if retcode != 0:
-        return False
-
-    info = json.loads(out)
-    try:
-        return (
-            # docker:
-            # https://docs.docker.com/reference/api/engine/version/v1.48/#tag/System/operation/SystemInfo
-            'name=rootless' in (info.get('SecurityOptions') or ()) or
-            # podman:
-            # https://docs.podman.io/en/latest/_static/api.html?version=v5.4#tag/system/operation/SystemInfoLibpod
-            info['host']['security']['rootless']
-        )
-    except KeyError:
-        return False
-
-
-def get_docker_user() -> tuple[str, ...]:  # pragma: win32 no cover
-    if _is_rootless():
-        return ()
-
+def get_docker_user() -> Tuple[str, ...]:  # pragma: win32 no cover
     try:
         return ('-u', f'{os.getuid()}:{os.getgid()}')
     except AttributeError:
         return ()
 
 
-def get_docker_tty(*, color: bool) -> tuple[str, ...]:  # pragma: win32 no cover  # noqa: E501
-    return (('--tty',) if color else ())
-
-
-def docker_cmd(*, color: bool) -> tuple[str, ...]:  # pragma: win32 no cover
+def docker_cmd() -> Tuple[str, ...]:  # pragma: win32 no cover
     return (
         'docker', 'run',
         '--rm',
-        *get_docker_tty(color=color),
         *get_docker_user(),
         # https://docs.docker.com/engine/reference/commandline/run/#mount-volumes-from-container-volumes-from
         # The `Z` option tells Docker to label the content with a private
@@ -157,25 +126,16 @@ def docker_cmd(*, color: bool) -> tuple[str, ...]:  # pragma: win32 no cover
 
 
 def run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
+        hook: Hook,
         file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
         color: bool,
-) -> tuple[int, bytes]:  # pragma: win32 no cover
+) -> Tuple[int, bytes]:  # pragma: win32 no cover
     # Rebuild the docker image in case it has gone missing, as many people do
     # automated cleanup of docker images.
-    build_docker_image(prefix, pull=False)
+    build_docker_image(hook.prefix, pull=False)
 
-    entry_exe, *cmd_rest = lang_base.hook_cmd(entry, args)
+    entry_exe, *cmd_rest = hook.cmd
 
-    entry_tag = ('--entrypoint', entry_exe, docker_tag(prefix))
-    return lang_base.run_xargs(
-        (*docker_cmd(color=color), *entry_tag, *cmd_rest),
-        file_args,
-        require_serial=require_serial,
-        color=color,
-    )
+    entry_tag = ('--entrypoint', entry_exe, docker_tag(hook.prefix))
+    cmd = (*docker_cmd(), *entry_tag, *cmd_rest)
+    return helpers.run_xargs(hook, cmd, file_args, color=color)

pre_commit/languages/docker_image.py

@@ -1,32 +1,20 @@
-from __future__ import annotations
+from typing import Sequence
+from typing import Tuple
 
-from collections.abc import Sequence
-
-from pre_commit import lang_base
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.languages.docker import docker_cmd
-from pre_commit.prefix import Prefix
 
 ENVIRONMENT_DIR = None
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-install_environment = lang_base.no_install
-in_env = lang_base.no_env
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
+install_environment = helpers.no_install
 
 
 def run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
+        hook: Hook,
         file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
         color: bool,
-) -> tuple[int, bytes]:  # pragma: win32 no cover
-    cmd = docker_cmd(color=color) + lang_base.hook_cmd(entry, args)
-    return lang_base.run_xargs(
-        cmd,
-        file_args,
-        require_serial=require_serial,
-        color=color,
-    )
+) -> Tuple[int, bytes]:  # pragma: win32 no cover
+    cmd = docker_cmd() + hook.cmd
+    return helpers.run_xargs(hook, cmd, file_args, color=color)

pre_commit/languages/dotnet.py

@@ -1,26 +1,23 @@
-from __future__ import annotations
-
 import contextlib
 import os.path
-import re
-import tempfile
-import xml.etree.ElementTree
-import zipfile
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
+import pre_commit.constants as C
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 
 ENVIRONMENT_DIR = 'dotnetenv'
 BIN_DIR = 'bin'
 
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
 def get_env_patch(venv: str) -> PatchesT:
@@ -30,82 +27,63 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+def in_env(prefix: Prefix) -> Generator[None, None, None]:
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT)
+    envdir = prefix.path(directory)
     with envcontext(get_env_patch(envdir)):
         yield
 
 
-@contextlib.contextmanager
-def _nuget_config_no_sources() -> Generator[str]:
-    with tempfile.TemporaryDirectory() as tmpdir:
-        nuget_config = os.path.join(tmpdir, 'nuget.config')
-        with open(nuget_config, 'w') as f:
-            f.write(
-                '<?xml version="1.0" encoding="utf-8"?>'
-                '<configuration>'
-                '  <packageSources>'
-                '    <clear />'
-                '  </packageSources>'
-                '</configuration>',
-            )
-        yield nuget_config
-
-
 def install_environment(
         prefix: Prefix,
         version: str,
         additional_dependencies: Sequence[str],
 ) -> None:
-    lang_base.assert_version_default('dotnet', version)
-    lang_base.assert_no_additional_deps('dotnet', additional_dependencies)
+    helpers.assert_version_default('dotnet', version)
+    helpers.assert_no_additional_deps('dotnet', additional_dependencies)
 
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    build_dir = prefix.path('pre-commit-build')
+    envdir = prefix.path(helpers.environment_dir(ENVIRONMENT_DIR, version))
+    with clean_path_on_failure(envdir):
+        build_dir = 'pre-commit-build'
 
-    # Build & pack nupkg file
-    lang_base.setup_cmd(
-        prefix,
-        (
-            'dotnet', 'pack',
-            '--configuration', 'Release',
-            '--property', f'PackageOutputPath={build_dir}',
-        ),
-    )
+        # Build & pack nupkg file
+        helpers.run_setup_cmd(
+            prefix,
+            (
+                'dotnet', 'pack',
+                '--configuration', 'Release',
+                '--output', build_dir,
+            ),
+        )
 
-    nupkg_dir = prefix.path(build_dir)
-    nupkgs = [x for x in os.listdir(nupkg_dir) if x.endswith('.nupkg')]
-
-    if not nupkgs:
-        raise AssertionError('could not find any build outputs to install')
-
-    for nupkg in nupkgs:
-        with zipfile.ZipFile(os.path.join(nupkg_dir, nupkg)) as f:
-            nuspec, = (x for x in f.namelist() if x.endswith('.nuspec'))
-            with f.open(nuspec) as spec:
-                tree = xml.etree.ElementTree.parse(spec)
-
-        namespace = re.match(r'{.*}', tree.getroot().tag)
-        if not namespace:
-            raise AssertionError('could not parse namespace from nuspec')
-
-        tool_id_element = tree.find(f'.//{namespace[0]}id')
-        if tool_id_element is None:
-            raise AssertionError('expected to find an "id" element')
-
-        tool_id = tool_id_element.text
-        if not tool_id:
-            raise AssertionError('"id" element missing tool name')
+        # Determine tool from the packaged file <tool_name>.<version>.nupkg
+        build_outputs = os.listdir(os.path.join(prefix.prefix_dir, build_dir))
+        if len(build_outputs) != 1:
+            raise NotImplementedError(
+                f"Can't handle multiple build outputs. Got {build_outputs}",
+            )
+        tool_name = build_outputs[0].split('.')[0]
 
         # Install to bin dir
-        with _nuget_config_no_sources() as nuget_config:
-            lang_base.setup_cmd(
-                prefix,
-                (
-                    'dotnet', 'tool', 'install',
-                    '--configfile', nuget_config,
-                    '--tool-path', os.path.join(envdir, BIN_DIR),
-                    '--add-source', build_dir,
-                    tool_id,
-                ),
-            )
+        helpers.run_setup_cmd(
+            prefix,
+            (
+                'dotnet', 'tool', 'install',
+                '--tool-path', os.path.join(envdir, BIN_DIR),
+                '--add-source', build_dir,
+                tool_name,
+            ),
+        )
+
+        # Clean the git dir, ignoring the environment dir
+        clean_cmd = ('git', 'clean', '-ffxd', '-e', f'{ENVIRONMENT_DIR}-*')
+        helpers.run_setup_cmd(prefix, clean_cmd)
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/fail.py

@@ -1,27 +1,20 @@
-from __future__ import annotations
+from typing import Sequence
+from typing import Tuple
 
-from collections.abc import Sequence
-
-from pre_commit import lang_base
-from pre_commit.prefix import Prefix
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 
 ENVIRONMENT_DIR = None
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-install_environment = lang_base.no_install
-in_env = lang_base.no_env
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
+install_environment = helpers.no_install
 
 
 def run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
+        hook: Hook,
         file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
         color: bool,
-) -> tuple[int, bytes]:
-    out = f'{entry}\n\n'.encode()
+) -> Tuple[int, bytes]:
+    out = f'{hook.entry}\n\n'.encode()
     out += b'\n'.join(f.encode() for f in file_args) + b'\n'
     return 1, out

pre_commit/languages/golang.py

@@ -1,161 +1,98 @@
-from __future__ import annotations
-
 import contextlib
-import functools
-import json
 import os.path
-import platform
-import shutil
 import sys
-import tarfile
-import tempfile
-import urllib.error
-import urllib.request
-import zipfile
-from collections.abc import Generator
-from collections.abc import Sequence
-from typing import ContextManager
-from typing import IO
-from typing import Protocol
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
 import pre_commit.constants as C
-from pre_commit import lang_base
+from pre_commit import git
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import Var
-from pre_commit.git import no_git_env
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output
+from pre_commit.util import cmd_output_b
 from pre_commit.util import rmtree
 
 ENVIRONMENT_DIR = 'golangenv'
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
-
-_ARCH_ALIASES = {
-    'x86_64': 'amd64',
-    'i386': '386',
-    'aarch64': 'arm64',
-    'armv8': 'arm64',
-    'armv7l': 'armv6l',
-}
-_ARCH = platform.machine().lower()
-_ARCH = _ARCH_ALIASES.get(_ARCH, _ARCH)
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
-class ExtractAll(Protocol):
-    def extractall(self, path: str) -> None: ...
-
-
-if sys.platform == 'win32':  # pragma: win32 cover
-    _EXT = 'zip'
-
-    def _open_archive(bio: IO[bytes]) -> ContextManager[ExtractAll]:
-        return zipfile.ZipFile(bio)
-else:  # pragma: win32 no cover
-    _EXT = 'tar.gz'
-
-    def _open_archive(bio: IO[bytes]) -> ContextManager[ExtractAll]:
-        return tarfile.open(fileobj=bio)
-
-
-@functools.lru_cache(maxsize=1)
-def get_default_version() -> str:
-    if lang_base.exe_exists('go'):
-        return 'system'
-    else:
-        return C.DEFAULT
-
-
-def get_env_patch(venv: str, version: str) -> PatchesT:
-    if version == 'system':
-        return (
-            ('PATH', (os.path.join(venv, 'bin'), os.pathsep, Var('PATH'))),
-        )
-
+def get_env_patch(venv: str) -> PatchesT:
     return (
-        ('GOROOT', os.path.join(venv, '.go')),
-        ('GOTOOLCHAIN', 'local'),
-        (
-            'PATH', (
-                os.path.join(venv, 'bin'), os.pathsep,
-                os.path.join(venv, '.go', 'bin'), os.pathsep, Var('PATH'),
-            ),
-        ),
+        ('PATH', (os.path.join(venv, 'bin'), os.pathsep, Var('PATH'))),
     )
 
 
-@functools.lru_cache
-def _infer_go_version(version: str) -> str:
-    if version != C.DEFAULT:
-        return version
-    resp = urllib.request.urlopen('https://go.dev/dl/?mode=json')
-    return json.load(resp)[0]['version'].removeprefix('go')
-
-
-def _get_url(version: str) -> str:
-    os_name = platform.system().lower()
-    version = _infer_go_version(version)
-    return f'https://dl.google.com/go/go{version}.{os_name}-{_ARCH}.{_EXT}'
-
-
-def _install_go(version: str, dest: str) -> None:
-    try:
-        resp = urllib.request.urlopen(_get_url(version))
-    except urllib.error.HTTPError as e:  # pragma: no cover
-        if e.code == 404:
-            raise ValueError(
-                f'Could not find a version matching your system requirements '
-                f'(os={platform.system().lower()}; arch={_ARCH})',
-            ) from e
-        else:
-            raise
-    else:
-        with tempfile.TemporaryFile() as f:
-            shutil.copyfileobj(resp, f)
-            f.seek(0)
-
-            with _open_archive(f) as archive:
-                archive.extractall(dest)
-        shutil.move(os.path.join(dest, 'go'), os.path.join(dest, '.go'))
-
-
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir, version)):
+def in_env(prefix: Prefix) -> Generator[None, None, None]:
+    envdir = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT),
+    )
+    with envcontext(get_env_patch(envdir)):
         yield
 
 
+def guess_go_dir(remote_url: str) -> str:
+    if remote_url.endswith('.git'):
+        remote_url = remote_url[:-1 * len('.git')]
+    looks_like_url = (
+        not remote_url.startswith('file://') and
+        ('//' in remote_url or '@' in remote_url)
+    )
+    remote_url = remote_url.replace(':', '/')
+    if looks_like_url:
+        _, _, remote_url = remote_url.rpartition('//')
+        _, _, remote_url = remote_url.rpartition('@')
+        return remote_url
+    else:
+        return 'unknown_src_dir'
+
+
 def install_environment(
         prefix: Prefix,
         version: str,
         additional_dependencies: Sequence[str],
 ) -> None:
-    env_dir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    helpers.assert_version_default('golang', version)
+    directory = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT),
+    )
 
-    if version != 'system':
-        _install_go(version, env_dir)
+    with clean_path_on_failure(directory):
+        remote = git.get_remote_url(prefix.prefix_dir)
+        repo_src_dir = os.path.join(directory, 'src', guess_go_dir(remote))
 
-    if sys.platform == 'cygwin':  # pragma: no cover
-        gopath = cmd_output('cygpath', '-w', env_dir)[1].strip()
-    else:
-        gopath = env_dir
+        # Clone into the goenv we'll create
+        cmd = ('git', 'clone', '--recursive', '.', repo_src_dir)
+        helpers.run_setup_cmd(prefix, cmd)
 
-    env = no_git_env(dict(os.environ, GOPATH=gopath))
-    env.pop('GOBIN', None)
-    if version != 'system':
-        env['GOTOOLCHAIN'] = 'local'
-        env['GOROOT'] = os.path.join(env_dir, '.go')
-        env['PATH'] = os.pathsep.join((
-            os.path.join(env_dir, '.go', 'bin'), os.environ['PATH'],
-        ))
+        if sys.platform == 'cygwin':  # pragma: no cover
+            _, gopath, _ = cmd_output('cygpath', '-w', directory)
+            gopath = gopath.strip()
+        else:
+            gopath = directory
+        env = dict(os.environ, GOPATH=gopath)
+        env.pop('GOBIN', None)
+        cmd_output_b('go', 'get', './...', cwd=repo_src_dir, env=env)
+        for dependency in additional_dependencies:
+            cmd_output_b('go', 'get', dependency, cwd=repo_src_dir, env=env)
+        # Same some disk space, we don't need these after installation
+        rmtree(prefix.path(directory, 'src'))
+        pkgdir = prefix.path(directory, 'pkg')
+        if os.path.exists(pkgdir):  # pragma: no cover (go<1.10)
+            rmtree(pkgdir)
 
-    lang_base.setup_cmd(prefix, ('go', 'install', './...'), env=env)
-    for dependency in additional_dependencies:
-        lang_base.setup_cmd(prefix, ('go', 'install', dependency), env=env)
 
-    # save some disk space -- we don't need this after installation
-    pkgdir = os.path.join(env_dir, 'pkg')
-    if os.path.exists(pkgdir):  # pragma: no branch (always true on windows?)
-        rmtree(pkgdir)
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/haskell.py

@@ -1,56 +0,0 @@
-from __future__ import annotations
-
-import contextlib
-import os.path
-from collections.abc import Generator
-from collections.abc import Sequence
-
-from pre_commit import lang_base
-from pre_commit.envcontext import envcontext
-from pre_commit.envcontext import PatchesT
-from pre_commit.envcontext import Var
-from pre_commit.errors import FatalError
-from pre_commit.prefix import Prefix
-
-ENVIRONMENT_DIR = 'hs_env'
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
-
-
-def get_env_patch(target_dir: str) -> PatchesT:
-    bin_path = os.path.join(target_dir, 'bin')
-    return (('PATH', (bin_path, os.pathsep, Var('PATH'))),)
-
-
-@contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir)):
-        yield
-
-
-def install_environment(
-    prefix: Prefix,
-    version: str,
-    additional_dependencies: Sequence[str],
-) -> None:
-    lang_base.assert_version_default('haskell', version)
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-
-    pkgs = [*prefix.star('.cabal'), *additional_dependencies]
-    if not pkgs:
-        raise FatalError('Expected .cabal files or additional_dependencies')
-
-    bindir = os.path.join(envdir, 'bin')
-    os.makedirs(bindir, exist_ok=True)
-    lang_base.setup_cmd(prefix, ('cabal', 'update'))
-    lang_base.setup_cmd(
-        prefix,
-        (
-            'cabal', 'install',
-            '--install-method', 'copy',
-            '--installdir', bindir,
-            *pkgs,
-        ),
-    )

pre_commit/languages/helpers.py

@@ -0,0 +1,136 @@
+import multiprocessing
+import os
+import random
+import re
+from typing import Any
+from typing import List
+from typing import Optional
+from typing import overload
+from typing import Sequence
+from typing import Tuple
+from typing import TYPE_CHECKING
+
+import pre_commit.constants as C
+from pre_commit import parse_shebang
+from pre_commit.hook import Hook
+from pre_commit.prefix import Prefix
+from pre_commit.util import cmd_output_b
+from pre_commit.xargs import xargs
+
+if TYPE_CHECKING:
+    from typing import NoReturn
+
+FIXED_RANDOM_SEED = 1542676187
+
+SHIMS_RE = re.compile(r'[/\\]shims[/\\]')
+
+
+def exe_exists(exe: str) -> bool:
+    found = parse_shebang.find_executable(exe)
+    if found is None:  # exe exists
+        return False
+
+    homedir = os.path.expanduser('~')
+    try:
+        common: Optional[str] = os.path.commonpath((found, homedir))
+    except ValueError:  # on windows, different drives raises ValueError
+        common = None
+
+    return (
+        # it is not in a /shims/ directory
+        not SHIMS_RE.search(found) and
+        (
+            # the homedir is / (docker, service user, etc.)
+            os.path.dirname(homedir) == homedir or
+            # the exe is not contained in the home directory
+            common != homedir
+        )
+    )
+
+
+def run_setup_cmd(prefix: Prefix, cmd: Tuple[str, ...], **kwargs: Any) -> None:
+    cmd_output_b(*cmd, cwd=prefix.prefix_dir, **kwargs)
+
+
+@overload
+def environment_dir(d: None, language_version: str) -> None: ...
+@overload
+def environment_dir(d: str, language_version: str) -> str: ...
+
+
+def environment_dir(d: Optional[str], language_version: str) -> Optional[str]:
+    if d is None:
+        return None
+    else:
+        return f'{d}-{language_version}'
+
+
+def assert_version_default(binary: str, version: str) -> None:
+    if version != C.DEFAULT:
+        raise AssertionError(
+            f'For now, pre-commit requires system-installed {binary}',
+        )
+
+
+def assert_no_additional_deps(
+        lang: str,
+        additional_deps: Sequence[str],
+) -> None:
+    if additional_deps:
+        raise AssertionError(
+            f'For now, pre-commit does not support '
+            f'additional_dependencies for {lang}',
+        )
+
+
+def basic_get_default_version() -> str:
+    return C.DEFAULT
+
+
+def basic_healthy(prefix: Prefix, language_version: str) -> bool:
+    return True
+
+
+def no_install(
+        prefix: Prefix,
+        version: str,
+        additional_dependencies: Sequence[str],
+) -> 'NoReturn':
+    raise AssertionError('This type is not installable')
+
+
+def target_concurrency(hook: Hook) -> int:
+    if hook.require_serial or 'PRE_COMMIT_NO_CONCURRENCY' in os.environ:
+        return 1
+    else:
+        # Travis appears to have a bunch of CPUs, but we can't use them all.
+        if 'TRAVIS' in os.environ:
+            return 2
+        else:
+            try:
+                return multiprocessing.cpu_count()
+            except NotImplementedError:
+                return 1
+
+
+def _shuffled(seq: Sequence[str]) -> List[str]:
+    """Deterministically shuffle"""
+    fixed_random = random.Random()
+    fixed_random.seed(FIXED_RANDOM_SEED, version=1)
+
+    seq = list(seq)
+    fixed_random.shuffle(seq)
+    return seq
+
+
+def run_xargs(
+        hook: Hook,
+        cmd: Tuple[str, ...],
+        file_args: Sequence[str],
+        **kwargs: Any,
+) -> Tuple[int, bytes]:
+    # Shuffle the files so that they more evenly fill out the xargs partitions,
+    # but do it deterministically in case a hook cares about ordering.
+    file_args = _shuffled(file_args)
+    kwargs['target_concurrency'] = target_concurrency(hook)
+    return xargs(cmd, file_args, **kwargs)

pre_commit/languages/julia.py

@@ -1,133 +0,0 @@
-from __future__ import annotations
-
-import contextlib
-import os
-import shutil
-from collections.abc import Generator
-from collections.abc import Sequence
-
-from pre_commit import lang_base
-from pre_commit.envcontext import envcontext
-from pre_commit.envcontext import PatchesT
-from pre_commit.envcontext import UNSET
-from pre_commit.prefix import Prefix
-from pre_commit.util import cmd_output_b
-
-ENVIRONMENT_DIR = 'juliaenv'
-health_check = lang_base.basic_health_check
-get_default_version = lang_base.basic_get_default_version
-
-
-def run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
-        file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
-        color: bool,
-) -> tuple[int, bytes]:
-    # `entry` is a (hook-repo relative) file followed by (optional) args, e.g.
-    # `bin/id.jl` or `bin/hook.jl --arg1 --arg2` so we
-    # 1) shell parse it and join with args with hook_cmd
-    # 2) prepend the hooks prefix path to the first argument (the file), unless
-    #    it is a local script
-    # 3) prepend `julia` as the interpreter
-
-    cmd = lang_base.hook_cmd(entry, args)
-    script = cmd[0] if is_local else prefix.path(cmd[0])
-    cmd = ('julia', '--startup-file=no', script, *cmd[1:])
-    return lang_base.run_xargs(
-        cmd,
-        file_args,
-        require_serial=require_serial,
-        color=color,
-    )
-
-
-def get_env_patch(target_dir: str, version: str) -> PatchesT:
-    return (
-        ('JULIA_LOAD_PATH', target_dir),
-        # May be set, remove it to not interfer with LOAD_PATH
-        ('JULIA_PROJECT', UNSET),
-    )
-
-
-@contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir, version)):
-        yield
-
-
-def install_environment(
-        prefix: Prefix,
-        version: str,
-        additional_dependencies: Sequence[str],
-) -> None:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with in_env(prefix, version):
-        # TODO: Support language_version with juliaup similar to rust via
-        # rustup
-        # if version != 'system':
-        #     ...
-
-        # Copy Project.toml to hook env if it exist
-        os.makedirs(envdir, exist_ok=True)
-        project_names = ('JuliaProject.toml', 'Project.toml')
-        project_found = False
-        for project_name in project_names:
-            project_file = prefix.path(project_name)
-            if not os.path.isfile(project_file):
-                continue
-            shutil.copy(project_file, envdir)
-            project_found = True
-            break
-
-        # If no project file was found we create an empty one so that the
-        # package manager doesn't error
-        if not project_found:
-            open(os.path.join(envdir, 'Project.toml'), 'a').close()
-
-        # Copy Manifest.toml to hook env if it exists
-        manifest_names = ('JuliaManifest.toml', 'Manifest.toml')
-        for manifest_name in manifest_names:
-            manifest_file = prefix.path(manifest_name)
-            if not os.path.isfile(manifest_file):
-                continue
-            shutil.copy(manifest_file, envdir)
-            break
-
-        # Julia code to instantiate the hook environment
-        julia_code = """
-        @assert length(ARGS) > 0
-        hook_env = ARGS[1]
-        deps = join(ARGS[2:end], " ")
-
-        # We prepend @stdlib here so that we can load the package manager even
-        # though `get_env_patch` limits `JULIA_LOAD_PATH` to just the hook env.
-        pushfirst!(LOAD_PATH, "@stdlib")
-        using Pkg
-        popfirst!(LOAD_PATH)
-
-        # Instantiate the environment shipped with the hook repo. If we have
-        # additional dependencies we disable precompilation in this step to
-        # avoid double work.
-        precompile = isempty(deps) ? "1" : "0"
-        withenv("JULIA_PKG_PRECOMPILE_AUTO" => precompile) do
-            Pkg.instantiate()
-        end
-
-        # Add additional dependencies (with precompilation)
-        if !isempty(deps)
-            withenv("JULIA_PKG_PRECOMPILE_AUTO" => "1") do
-                Pkg.REPLMode.pkgstr("add " * deps)
-            end
-        end
-        """
-        cmd_output_b(
-            'julia', '--startup-file=no', '-e', julia_code, '--', envdir,
-            *additional_dependencies,
-            cwd=prefix.prefix_dir,
-        )

pre_commit/languages/lua.py

@@ -1,75 +0,0 @@
-from __future__ import annotations
-
-import contextlib
-import os
-import sys
-from collections.abc import Generator
-from collections.abc import Sequence
-
-from pre_commit import lang_base
-from pre_commit.envcontext import envcontext
-from pre_commit.envcontext import PatchesT
-from pre_commit.envcontext import Var
-from pre_commit.prefix import Prefix
-from pre_commit.util import cmd_output
-
-ENVIRONMENT_DIR = 'lua_env'
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
-
-
-def _get_lua_version() -> str:  # pragma: win32 no cover
-    """Get the Lua version used in file paths."""
-    _, stdout, _ = cmd_output('luarocks', 'config', '--lua-ver')
-    return stdout.strip()
-
-
-def get_env_patch(d: str) -> PatchesT:  # pragma: win32 no cover
-    version = _get_lua_version()
-    so_ext = 'dll' if sys.platform == 'win32' else 'so'
-    return (
-        ('PATH', (os.path.join(d, 'bin'), os.pathsep, Var('PATH'))),
-        (
-            'LUA_PATH', (
-                os.path.join(d, 'share', 'lua', version, '?.lua;'),
-                os.path.join(d, 'share', 'lua', version, '?', 'init.lua;;'),
-            ),
-        ),
-        (
-            'LUA_CPATH',
-            (os.path.join(d, 'lib', 'lua', version, f'?.{so_ext};;'),),
-        ),
-    )
-
-
-@contextlib.contextmanager  # pragma: win32 no cover
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir)):
-        yield
-
-
-def install_environment(
-    prefix: Prefix,
-    version: str,
-    additional_dependencies: Sequence[str],
-) -> None:  # pragma: win32 no cover
-    lang_base.assert_version_default('lua', version)
-
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with in_env(prefix, version):
-        # luarocks doesn't bootstrap a tree prior to installing
-        # so ensure the directory exists.
-        os.makedirs(envdir, exist_ok=True)
-
-        # Older luarocks (e.g., 2.4.2) expect the rockspec as an arg
-        for rockspec in prefix.star('.rockspec'):
-            make_cmd = ('luarocks', '--tree', envdir, 'make', rockspec)
-            lang_base.setup_cmd(prefix, make_cmd)
-
-        # luarocks can't install multiple packages at once
-        # so install them individually.
-        for dependency in additional_dependencies:
-            cmd = ('luarocks', '--tree', envdir, 'install', dependency)
-            lang_base.setup_cmd(prefix, cmd)

pre_commit/languages/node.py

@@ -1,26 +1,26 @@
-from __future__ import annotations
-
 import contextlib
 import functools
 import os
 import sys
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
 import pre_commit.constants as C
-from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import UNSET
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.languages.python import bin_dir
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output
 from pre_commit.util import cmd_output_b
 from pre_commit.util import rmtree
 
 ENVIRONMENT_DIR = 'node_env'
-run_hook = lang_base.basic_run_hook
 
 
 @functools.lru_cache(maxsize=1)
@@ -30,12 +30,17 @@ def get_default_version() -> str:
         return C.DEFAULT
     # if node is already installed, we can save a bunch of setup time by
     # using the installed version
-    elif all(lang_base.exe_exists(exe) for exe in ('node', 'npm')):
+    elif all(helpers.exe_exists(exe) for exe in ('node', 'npm')):
         return 'system'
     else:
         return C.DEFAULT
 
 
+def _envdir(prefix: Prefix, version: str) -> str:
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, version)
+    return prefix.path(directory)
+
+
 def get_env_patch(venv: str) -> PatchesT:
     if sys.platform == 'cygwin':  # pragma: no cover
         _, win_venv, _ = cmd_output('cygpath', '-w', venv)
@@ -59,52 +64,64 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir)):
+def in_env(
+        prefix: Prefix,
+        language_version: str,
+) -> Generator[None, None, None]:
+    with envcontext(get_env_patch(_envdir(prefix, language_version))):
         yield
 
 
-def health_check(prefix: Prefix, version: str) -> str | None:
-    with in_env(prefix, version):
-        retcode, _, _ = cmd_output_b('node', '--version', check=False)
-        if retcode != 0:  # pragma: win32 no cover
-            return f'`node --version` returned {retcode}'
-        else:
-            return None
+def healthy(prefix: Prefix, language_version: str) -> bool:
+    with in_env(prefix, language_version):
+        retcode, _, _ = cmd_output_b('node', '--version', retcode=None)
+        return retcode == 0
 
 
 def install_environment(
         prefix: Prefix, version: str, additional_dependencies: Sequence[str],
 ) -> None:
+    additional_dependencies = tuple(additional_dependencies)
     assert prefix.exists('package.json')
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    envdir = _envdir(prefix, version)
 
     # https://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx?f=255&MSPPError=-2147217396#maxpath
     if sys.platform == 'win32':  # pragma: no cover
         envdir = fr'\\?\{os.path.normpath(envdir)}'
-    cmd = [sys.executable, '-mnodeenv', '--prebuilt', '--clean-src', envdir]
-    if version != C.DEFAULT:
-        cmd.extend(['-n', version])
-    cmd_output_b(*cmd)
+    with clean_path_on_failure(envdir):
+        cmd = [
+            sys.executable, '-mnodeenv', '--prebuilt', '--clean-src', envdir,
+        ]
+        if version != C.DEFAULT:
+            cmd.extend(['-n', version])
+        cmd_output_b(*cmd)
 
-    with in_env(prefix, version):
-        # https://npm.community/t/npm-install-g-git-vs-git-clone-cd-npm-install-g/5449
-        # install as if we installed from git
+        with in_env(prefix, version):
+            # https://npm.community/t/npm-install-g-git-vs-git-clone-cd-npm-install-g/5449
+            # install as if we installed from git
 
-        local_install_cmd = (
-            'npm', 'install', '--include=dev', '--include=prod',
-            '--ignore-prepublish', '--no-progress', '--no-save',
-        )
-        lang_base.setup_cmd(prefix, local_install_cmd)
+            local_install_cmd = (
+                'npm', 'install', '--dev', '--prod',
+                '--ignore-prepublish', '--no-progress', '--no-save',
+            )
+            helpers.run_setup_cmd(prefix, local_install_cmd)
 
-        _, pkg, _ = cmd_output('npm', 'pack', cwd=prefix.prefix_dir)
-        pkg = prefix.path(pkg.strip())
+            _, pkg, _ = cmd_output('npm', 'pack', cwd=prefix.prefix_dir)
+            pkg = prefix.path(pkg.strip())
 
-        install = ('npm', 'install', '-g', pkg, *additional_dependencies)
-        lang_base.setup_cmd(prefix, install)
+            install = ('npm', 'install', '-g', pkg, *additional_dependencies)
+            helpers.run_setup_cmd(prefix, install)
 
-        # clean these up after installation
-        if prefix.exists('node_modules'):  # pragma: win32 no cover
-            rmtree(prefix.path('node_modules'))
-        os.remove(pkg)
+            # clean these up after installation
+            if prefix.exists('node_modules'):  # pragma: win32 no cover
+                rmtree(prefix.path('node_modules'))
+            os.remove(pkg)
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix, hook.language_version):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/perl.py

@@ -1,21 +1,26 @@
-from __future__ import annotations
-
 import contextlib
 import os
 import shlex
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 
 ENVIRONMENT_DIR = 'perl_env'
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
+
+
+def _envdir(prefix: Prefix, version: str) -> str:
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, version)
+    return prefix.path(directory)
 
 
 def get_env_patch(venv: str) -> PatchesT:
@@ -33,18 +38,30 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir)):
+def in_env(
+        prefix: Prefix,
+        language_version: str,
+) -> Generator[None, None, None]:
+    with envcontext(get_env_patch(_envdir(prefix, language_version))):
         yield
 
 
 def install_environment(
         prefix: Prefix, version: str, additional_dependencies: Sequence[str],
 ) -> None:
-    lang_base.assert_version_default('perl', version)
+    helpers.assert_version_default('perl', version)
 
-    with in_env(prefix, version):
-        lang_base.setup_cmd(
-            prefix, ('cpan', '-T', '.', *additional_dependencies),
-        )
+    with clean_path_on_failure(_envdir(prefix, version)):
+        with in_env(prefix, version):
+            helpers.run_setup_cmd(
+                prefix, ('cpan', '-T', '.', *additional_dependencies),
+            )
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix, hook.language_version):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/pygrep.py

@@ -1,22 +1,21 @@
-from __future__ import annotations
-
 import argparse
 import re
 import sys
-from collections.abc import Sequence
-from re import Pattern
 from typing import NamedTuple
+from typing import Optional
+from typing import Pattern
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
 from pre_commit import output
-from pre_commit.prefix import Prefix
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.xargs import xargs
 
 ENVIRONMENT_DIR = None
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-install_environment = lang_base.no_install
-in_env = lang_base.no_env
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
+install_environment = helpers.no_install
 
 
 def _process_filename_by_line(pattern: Pattern[bytes], filename: str) -> int:
@@ -88,20 +87,15 @@ FNS = {
 
 
 def run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
+        hook: Hook,
         file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
         color: bool,
-) -> tuple[int, bytes]:
-    cmd = (sys.executable, '-m', __name__, *args, entry)
-    return xargs(cmd, file_args, color=color)
+) -> Tuple[int, bytes]:
+    exe = (sys.executable, '-m', __name__) + tuple(hook.args) + (hook.entry,)
+    return xargs(exe, file_args, color=color)
 
 
-def main(argv: Sequence[str] | None = None) -> int:
+def main(argv: Optional[Sequence[str]] = None) -> int:
     parser = argparse.ArgumentParser(
         description=(
             'grep-like finder using python regexes.  Unlike grep, this tool '
@@ -130,4 +124,4 @@ def main(argv: Sequence[str] | None = None) -> int:
 
 
 if __name__ == '__main__':
-    raise SystemExit(main())
+    exit(main())

pre_commit/languages/python.py

@@ -1,30 +1,32 @@
-from __future__ import annotations
-
 import contextlib
 import functools
 import os
 import sys
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Dict
+from typing import Generator
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
 
 import pre_commit.constants as C
-from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import UNSET
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.parse_shebang import find_executable
 from pre_commit.prefix import Prefix
 from pre_commit.util import CalledProcessError
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output
 from pre_commit.util import cmd_output_b
 from pre_commit.util import win_exe
 
 ENVIRONMENT_DIR = 'py_env'
-run_hook = lang_base.basic_run_hook
 
 
-@functools.cache
+@functools.lru_cache(maxsize=None)
 def _version_info(exe: str) -> str:
     prog = 'import sys;print(".".join(str(p) for p in sys.version_info))'
     try:
@@ -33,7 +35,7 @@ def _version_info(exe: str) -> str:
         return f'<<error retrieving version from {exe}>>'
 
 
-def _read_pyvenv_cfg(filename: str) -> dict[str, str]:
+def _read_pyvenv_cfg(filename: str) -> Dict[str, str]:
     ret = {}
     with open(filename, encoding='UTF-8') as f:
         for line in f:
@@ -48,7 +50,7 @@ def _read_pyvenv_cfg(filename: str) -> dict[str, str]:
 
 def bin_dir(venv: str) -> str:
     """On windows there's a different directory for the virtualenv"""
-    bin_part = 'Scripts' if sys.platform == 'win32' else 'bin'
+    bin_part = 'Scripts' if os.name == 'nt' else 'bin'
     return os.path.join(venv, bin_part)
 
 
@@ -63,9 +65,9 @@ def get_env_patch(venv: str) -> PatchesT:
 
 def _find_by_py_launcher(
         version: str,
-) -> str | None:  # pragma: no cover (windows only)
+) -> Optional[str]:  # pragma: no cover (windows only)
     if version.startswith('python'):
-        num = version.removeprefix('python')
+        num = version[len('python'):]
         cmd = ('py', f'-{num}', '-c', 'import sys; print(sys.executable)')
         env = dict(os.environ, PYTHONIOENCODING='UTF-8')
         try:
@@ -75,15 +77,8 @@ def _find_by_py_launcher(
     return None
 
 
-def _impl_exe_name() -> str:
-    if sys.implementation.name == 'cpython':  # pragma: cpython cover
-        return 'python'
-    else:  # pragma: cpython no cover
-        return sys.implementation.name  # pypy mostly
-
-
-def _find_by_sys_executable() -> str | None:
-    def _norm(path: str) -> str | None:
+def _find_by_sys_executable() -> Optional[str]:
+    def _norm(path: str) -> Optional[str]:
         _, exe = os.path.split(path.lower())
         exe, _, _ = exe.partition('.exe')
         if exe not in {'python', 'pythonw'} and find_executable(exe):
@@ -107,25 +102,18 @@ def _find_by_sys_executable() -> str | None:
 
 @functools.lru_cache(maxsize=1)
 def get_default_version() -> str:  # pragma: no cover (platform dependent)
-    v_major = f'{sys.version_info[0]}'
-    v_minor = f'{sys.version_info[0]}.{sys.version_info[1]}'
+    # First attempt from `sys.executable` (or the realpath)
+    exe = _find_by_sys_executable()
+    if exe:
+        return exe
 
-    # attempt the likely implementation exe
-    for potential in (v_minor, v_major):
-        exe = f'{_impl_exe_name()}{potential}'
-        if find_executable(exe):
-            return exe
+    # Next try the `pythonX.X` executable
+    exe = f'python{sys.version_info[0]}.{sys.version_info[1]}'
+    if find_executable(exe):
+        return exe
 
-    # next try `sys.executable` (or the realpath)
-    maybe_exe = _find_by_sys_executable()
-    if maybe_exe:
-        return maybe_exe
-
-    # maybe on windows we can find it via py launcher?
-    if sys.platform == 'win32':  # pragma: win32 cover
-        exe = f'python{v_minor}'
-        if _find_by_py_launcher(exe):
-            return exe
+    if _find_by_py_launcher(exe):
+        return exe
 
     # We tried!
     return C.DEFAULT
@@ -138,20 +126,20 @@ def _sys_executable_matches(version: str) -> bool:
         return False
 
     try:
-        info = tuple(int(p) for p in version.removeprefix('python').split('.'))
+        info = tuple(int(p) for p in version[len('python'):].split('.'))
     except ValueError:
         return False
 
     return sys.version_info[:len(info)] == info
 
 
-def norm_version(version: str) -> str | None:
+def norm_version(version: str) -> Optional[str]:
     if version == C.DEFAULT:  # use virtualenv's default
         return None
     elif _sys_executable_matches(version):  # virtualenv defaults to our exe
         return None
 
-    if sys.platform == 'win32':  # pragma: no cover (windows)
+    if os.name == 'nt':  # pragma: no cover (windows)
         version_exec = _find_by_py_launcher(version)
         if version_exec:
             return version_exec
@@ -166,49 +154,37 @@ def norm_version(version: str) -> str | None:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+def in_env(
+        prefix: Prefix,
+        language_version: str,
+) -> Generator[None, None, None]:
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, language_version)
+    envdir = prefix.path(directory)
     with envcontext(get_env_patch(envdir)):
         yield
 
 
-def health_check(prefix: Prefix, version: str) -> str | None:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+def healthy(prefix: Prefix, language_version: str) -> bool:
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, language_version)
+    envdir = prefix.path(directory)
     pyvenv_cfg = os.path.join(envdir, 'pyvenv.cfg')
 
     # created with "old" virtualenv
     if not os.path.exists(pyvenv_cfg):
-        return 'pyvenv.cfg does not exist (old virtualenv?)'
+        return False
 
     exe_name = win_exe('python')
     py_exe = prefix.path(bin_dir(envdir), exe_name)
     cfg = _read_pyvenv_cfg(pyvenv_cfg)
 
-    if 'version_info' not in cfg:
-        return "created virtualenv's pyvenv.cfg is missing `version_info`"
-
-    # always use uncached lookup here in case we replaced an unhealthy env
-    virtualenv_version = _version_info.__wrapped__(py_exe)
-    if virtualenv_version != cfg['version_info']:
-        return (
-            f'virtualenv python version did not match created version:\n'
-            f'- actual version: {virtualenv_version}\n'
-            f'- expected version: {cfg["version_info"]}\n'
+    return (
+        'version_info' in cfg and
+        # always use uncached lookup here in case we replaced an unhealthy env
+        _version_info.__wrapped__(py_exe) == cfg['version_info'] and (
+            'base-executable' not in cfg or
+            _version_info(cfg['base-executable']) == cfg['version_info']
         )
-
-    # made with an older version of virtualenv? skip `base-executable` check
-    if 'base-executable' not in cfg:
-        return None
-
-    base_exe_version = _version_info(cfg['base-executable'])
-    if base_exe_version != cfg['version_info']:
-        return (
-            f'base executable python version does not match created version:\n'
-            f'- base-executable version: {base_exe_version}\n'
-            f'- expected version: {cfg["version_info"]}\n'
-        )
-    else:
-        return None
+    )
 
 
 def install_environment(
@@ -216,13 +192,23 @@ def install_environment(
         version: str,
         additional_dependencies: Sequence[str],
 ) -> None:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    envdir = prefix.path(helpers.environment_dir(ENVIRONMENT_DIR, version))
     venv_cmd = [sys.executable, '-mvirtualenv', envdir]
     python = norm_version(version)
     if python is not None:
         venv_cmd.extend(('-p', python))
     install_cmd = ('python', '-mpip', 'install', '.', *additional_dependencies)
 
-    cmd_output_b(*venv_cmd, cwd='/')
-    with in_env(prefix, version):
-        lang_base.setup_cmd(prefix, install_cmd)
+    with clean_path_on_failure(envdir):
+        cmd_output_b(*venv_cmd, cwd='/')
+        with in_env(prefix, version):
+            helpers.run_setup_cmd(prefix, install_cmd)
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix, hook.language_version):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/r.py

@@ -1,160 +1,61 @@
-from __future__ import annotations
-
 import contextlib
 import os
 import shlex
 import shutil
-import tempfile
-import textwrap
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
-from pre_commit.envcontext import UNSET
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
-from pre_commit.util import cmd_output
-from pre_commit.util import win_exe
+from pre_commit.util import clean_path_on_failure
+from pre_commit.util import cmd_output_b
 
 ENVIRONMENT_DIR = 'renv'
-get_default_version = lang_base.basic_get_default_version
-
-_RENV_ACTIVATED_OPTS = (
-    '--no-save', '--no-restore', '--no-site-file', '--no-environ',
-)
-
-
-def _execute_r(
-        code: str, *,
-        prefix: Prefix, version: str, args: Sequence[str] = (), cwd: str,
-        cli_opts: Sequence[str],
-) -> str:
-    with in_env(prefix, version), _r_code_in_tempfile(code) as f:
-        _, out, _ = cmd_output(
-            _rscript_exec(), *cli_opts, f, *args, cwd=cwd,
-        )
-    return out.rstrip('\n')
-
-
-def _execute_r_in_renv(
-        code: str, *,
-        prefix: Prefix, version: str, args: Sequence[str] = (), cwd: str,
-) -> str:
-    return _execute_r(
-        code=code, prefix=prefix, version=version, args=args, cwd=cwd,
-        cli_opts=_RENV_ACTIVATED_OPTS,
-    )
-
-
-def _execute_vanilla_r(
-        code: str, *,
-        prefix: Prefix, version: str, args: Sequence[str] = (), cwd: str,
-) -> str:
-    return _execute_r(
-        code=code, prefix=prefix, version=version, args=args, cwd=cwd,
-        cli_opts=('--vanilla',),
-    )
-
-
-def _read_installed_version(envdir: str, prefix: Prefix, version: str) -> str:
-    return _execute_r_in_renv(
-        'cat(renv::settings$r.version())',
-        prefix=prefix, version=version,
-        cwd=envdir,
-    )
-
-
-def _read_executable_version(envdir: str, prefix: Prefix, version: str) -> str:
-    return _execute_r_in_renv(
-        'cat(as.character(getRversion()))',
-        prefix=prefix, version=version,
-        cwd=envdir,
-    )
-
-
-def _write_current_r_version(
-        envdir: str, prefix: Prefix, version: str,
-) -> None:
-    _execute_r_in_renv(
-        'renv::settings$r.version(as.character(getRversion()))',
-        prefix=prefix, version=version,
-        cwd=envdir,
-    )
-
-
-def health_check(prefix: Prefix, version: str) -> str | None:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-
-    r_version_installation = _read_installed_version(
-        envdir=envdir, prefix=prefix, version=version,
-    )
-    r_version_current_executable = _read_executable_version(
-        envdir=envdir, prefix=prefix, version=version,
-    )
-    if r_version_installation in {'NULL', ''}:
-        return (
-            f'Hooks were installed with an unknown R version. R version for '
-            f'hook repo now set to {r_version_current_executable}'
-        )
-    elif r_version_installation != r_version_current_executable:
-        return (
-            f'Hooks were installed for R version {r_version_installation}, '
-            f'but current R executable has version '
-            f'{r_version_current_executable}'
-        )
-
-    return None
-
-
-@contextlib.contextmanager
-def _r_code_in_tempfile(code: str) -> Generator[str]:
-    """
-    To avoid quoting and escaping issues, avoid `Rscript [options] -e {expr}`
-    but use `Rscript [options] path/to/file_with_expr.R`
-    """
-    with tempfile.TemporaryDirectory() as tmpdir:
-        fname = os.path.join(tmpdir, 'script.R')
-        with open(fname, 'w') as f:
-            f.write(_inline_r_setup(textwrap.dedent(code)))
-        yield fname
+RSCRIPT_OPTS = ('--no-save', '--no-restore', '--no-site-file', '--no-environ')
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
 def get_env_patch(venv: str) -> PatchesT:
     return (
         ('R_PROFILE_USER', os.path.join(venv, 'activate.R')),
-        ('RENV_PROJECT', UNSET),
     )
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+def in_env(
+        prefix: Prefix,
+        language_version: str,
+) -> Generator[None, None, None]:
+    envdir = _get_env_dir(prefix, language_version)
     with envcontext(get_env_patch(envdir)):
         yield
 
 
-def _prefix_if_file_entry(
-        entry: list[str],
-        prefix: Prefix,
-        *,
-        is_local: bool,
+def _get_env_dir(prefix: Prefix, version: str) -> str:
+    return prefix.path(helpers.environment_dir(ENVIRONMENT_DIR, version))
+
+
+def _prefix_if_non_local_file_entry(
+    entry: Sequence[str],
+    prefix: Prefix,
+    src: str,
 ) -> Sequence[str]:
-    if entry[1] == '-e' or is_local:
+    if entry[1] == '-e':
         return entry[1:]
     else:
-        return (prefix.path(entry[1]),)
+        if src == 'local':
+            path = entry[1]
+        else:
+            path = prefix.path(entry[1])
+        return (path,)
 
 
-def _rscript_exec() -> str:
-    r_home = os.environ.get('R_HOME')
-    if r_home is None:
-        return 'Rscript'
-    else:
-        return os.path.join(r_home, 'bin', win_exe('Rscript'))
-
-
-def _entry_validate(entry: list[str]) -> None:
+def _entry_validate(entry: Sequence[str]) -> None:
     """
     Allowed entries:
     # Rscript -e expr
@@ -168,23 +69,20 @@ def _entry_validate(entry: list[str]) -> None:
             raise ValueError('You can supply at most one expression.')
     elif len(entry) > 2:
         raise ValueError(
-            'The only valid syntax is `Rscript -e {expr}`'
+            'The only valid syntax is `Rscript -e {expr}`',
             'or `Rscript path/to/hook/script`',
         )
 
 
-def _cmd_from_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
-        *,
-        is_local: bool,
-) -> tuple[str, ...]:
-    cmd = shlex.split(entry)
-    _entry_validate(cmd)
+def _cmd_from_hook(hook: Hook) -> Tuple[str, ...]:
+    entry = shlex.split(hook.entry)
+    _entry_validate(entry)
 
-    cmd_part = _prefix_if_file_entry(cmd, prefix, is_local=is_local)
-    return (cmd[0], *_RENV_ACTIVATED_OPTS, *cmd_part, *args)
+    return (
+        *entry[:1], *RSCRIPT_OPTS,
+        *_prefix_if_non_local_file_entry(entry, hook.prefix, hook.src),
+        *hook.args,
+    )
 
 
 def install_environment(
@@ -192,87 +90,59 @@ def install_environment(
         version: str,
         additional_dependencies: Sequence[str],
 ) -> None:
-    lang_base.assert_version_default('r', version)
-
-    env_dir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    os.makedirs(env_dir, exist_ok=True)
-    shutil.copy(prefix.path('renv.lock'), env_dir)
-    shutil.copytree(prefix.path('renv'), os.path.join(env_dir, 'renv'))
-
-    r_code_inst_environment = f"""\
-        prefix_dir <- {prefix.prefix_dir!r}
-        options(
-            repos = c(CRAN = "https://cran.rstudio.com"),
-            renv.consent = TRUE
-        )
-        source("renv/activate.R")
-        renv::restore()
-        activate_statement <- paste0(
-          'suppressWarnings({{',
-          'old <- setwd("', getwd(), '"); ',
-          'source("renv/activate.R"); ',
-          'setwd(old); ',
-          'renv::load("', getwd(), '");}})'
-        )
-        writeLines(activate_statement, 'activate.R')
-        is_package <- tryCatch(
-          {{
-              path_desc <- file.path(prefix_dir, 'DESCRIPTION')
-              suppressWarnings(desc <- read.dcf(path_desc))
-              "Package" %in% colnames(desc)
-          }},
-          error = function(...) FALSE
-        )
-        if (is_package) {{
-            renv::install(prefix_dir)
-        }}
-        """
-    _execute_vanilla_r(
-        r_code_inst_environment,
-        prefix=prefix, version=version, cwd=env_dir,
-    )
-
-    _write_current_r_version(envdir=env_dir, prefix=prefix, version=version)
-    if additional_dependencies:
-        r_code_inst_add = 'renv::install(commandArgs(trailingOnly = TRUE))'
-        _execute_r_in_renv(
-            code=r_code_inst_add, prefix=prefix, version=version,
-            args=additional_dependencies,
+    env_dir = _get_env_dir(prefix, version)
+    with clean_path_on_failure(env_dir):
+        os.makedirs(env_dir, exist_ok=True)
+        shutil.copy(prefix.path('renv.lock'), env_dir)
+        shutil.copytree(prefix.path('renv'), os.path.join(env_dir, 'renv'))
+        cmd_output_b(
+            'Rscript', '--vanilla', '-e',
+            f"""\
+            prefix_dir <- {prefix.prefix_dir!r}
+            options(
+                repos = c(CRAN = "https://cran.rstudio.com"),
+                renv.consent = TRUE
+            )
+            source("renv/activate.R")
+            renv::restore()
+            activate_statement <- paste0(
+              'suppressWarnings({{',
+              'old <- setwd("', getwd(), '"); ',
+              'source("renv/activate.R"); ',
+              'setwd(old); ',
+              'renv::load("', getwd(), '");}})'
+            )
+            writeLines(activate_statement, 'activate.R')
+            is_package <- tryCatch(
+              {{
+                  path_desc <- file.path(prefix_dir, 'DESCRIPTION')
+                  suppressWarnings(desc <- read.dcf(path_desc))
+                  "Package" %in% colnames(desc)
+              }},
+              error = function(...) FALSE
+            )
+            if (is_package) {{
+                renv::install(prefix_dir)
+            }}
+            """,
             cwd=env_dir,
         )
-
-
-def _inline_r_setup(code: str) -> str:
-    """
-    Some behaviour of R cannot be configured via env variables, but can
-    only be configured via R options once R has started. These are set here.
-    """
-    with_option = [
-        textwrap.dedent("""\
-        options(
-            install.packages.compile.from.source = "never",
-            pkgType = "binary"
-        )
-        """),
-        code,
-    ]
-    return '\n'.join(with_option)
+        if additional_dependencies:
+            with in_env(prefix, version):
+                cmd_output_b(
+                    'Rscript', *RSCRIPT_OPTS, '-e',
+                    'renv::install(commandArgs(trailingOnly = TRUE))',
+                    *additional_dependencies,
+                    cwd=env_dir,
+                )
 
 
 def run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
+        hook: Hook,
         file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
         color: bool,
-) -> tuple[int, bytes]:
-    cmd = _cmd_from_hook(prefix, entry, args, is_local=is_local)
-    return lang_base.run_xargs(
-        cmd,
-        file_args,
-        require_serial=require_serial,
-        color=color,
-    )
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix, hook.language_version):
+        return helpers.run_xargs(
+            hook, _cmd_from_hook(hook), file_args, color=color,
+        )

pre_commit/languages/ruby.py

@@ -1,37 +1,31 @@
-from __future__ import annotations
-
 import contextlib
 import functools
-import importlib.resources
 import os.path
 import shutil
 import tarfile
-from collections.abc import Generator
-from collections.abc import Sequence
-from typing import IO
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
 import pre_commit.constants as C
-from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import UNSET
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
 from pre_commit.util import CalledProcessError
+from pre_commit.util import clean_path_on_failure
+from pre_commit.util import resource_bytesio
 
 ENVIRONMENT_DIR = 'rbenv'
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
-
-
-def _resource_bytesio(filename: str) -> IO[bytes]:
-    files = importlib.resources.files('pre_commit.resources')
-    return files.joinpath(filename).open('rb')
+healthy = helpers.basic_healthy
 
 
 @functools.lru_cache(maxsize=1)
 def get_default_version() -> str:
-    if all(lang_base.exe_exists(exe) for exe in ('ruby', 'gem')):
+    if all(helpers.exe_exists(exe) for exe in ('ruby', 'gem')):
         return 'system'
     else:
         return C.DEFAULT
@@ -73,14 +67,19 @@ def get_env_patch(
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir, version)):
+def in_env(
+        prefix: Prefix,
+        language_version: str,
+) -> Generator[None, None, None]:
+    envdir = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, language_version),
+    )
+    with envcontext(get_env_patch(envdir, language_version)):
         yield
 
 
 def _extract_resource(filename: str, dest: str) -> None:
-    with _resource_bytesio(filename) as bio:
+    with resource_bytesio(filename) as bio:
         with tarfile.open(fileobj=bio) as tf:
             tf.extractall(dest)
 
@@ -89,14 +88,14 @@ def _install_rbenv(
         prefix: Prefix,
         version: str,
 ) -> None:  # pragma: win32 no cover
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, version)
 
     _extract_resource('rbenv.tar.gz', prefix.path('.'))
-    shutil.move(prefix.path('rbenv'), envdir)
+    shutil.move(prefix.path('rbenv'), prefix.path(directory))
 
     # Only install ruby-build if the version is specified
     if version != C.DEFAULT:
-        plugins_dir = os.path.join(envdir, 'plugins')
+        plugins_dir = prefix.path(directory, 'plugins')
         _extract_resource('ruby-download.tar.gz', plugins_dir)
         _extract_resource('ruby-build.tar.gz', plugins_dir)
 
@@ -106,40 +105,47 @@ def _install_ruby(
         version: str,
 ) -> None:  # pragma: win32 no cover
     try:
-        lang_base.setup_cmd(prefix, ('rbenv', 'download', version))
+        helpers.run_setup_cmd(prefix, ('rbenv', 'download', version))
     except CalledProcessError:  # pragma: no cover (usually find with download)
         # Failed to download from mirror for some reason, build it instead
-        lang_base.setup_cmd(prefix, ('rbenv', 'install', version))
+        helpers.run_setup_cmd(prefix, ('rbenv', 'install', version))
 
 
 def install_environment(
         prefix: Prefix, version: str, additional_dependencies: Sequence[str],
 ) -> None:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    additional_dependencies = tuple(additional_dependencies)
+    directory = helpers.environment_dir(ENVIRONMENT_DIR, version)
+    with clean_path_on_failure(prefix.path(directory)):
+        if version != 'system':  # pragma: win32 no cover
+            _install_rbenv(prefix, version)
+            with in_env(prefix, version):
+                # Need to call this before installing so rbenv's directories
+                # are set up
+                helpers.run_setup_cmd(prefix, ('rbenv', 'init', '-'))
+                if version != C.DEFAULT:
+                    _install_ruby(prefix, version)
+                # Need to call this after installing to set up the shims
+                helpers.run_setup_cmd(prefix, ('rbenv', 'rehash'))
 
-    if version != 'system':  # pragma: win32 no cover
-        _install_rbenv(prefix, version)
         with in_env(prefix, version):
-            # Need to call this before installing so rbenv's directories
-            # are set up
-            lang_base.setup_cmd(prefix, ('rbenv', 'init', '-'))
-            if version != C.DEFAULT:
-                _install_ruby(prefix, version)
-            # Need to call this after installing to set up the shims
-            lang_base.setup_cmd(prefix, ('rbenv', 'rehash'))
+            helpers.run_setup_cmd(
+                prefix, ('gem', 'build', *prefix.star('.gemspec')),
+            )
+            helpers.run_setup_cmd(
+                prefix,
+                (
+                    'gem', 'install',
+                    '--no-document', '--no-format-executable',
+                    *prefix.star('.gem'), *additional_dependencies,
+                ),
+            )
 
-    with in_env(prefix, version):
-        lang_base.setup_cmd(
-            prefix, ('gem', 'build', *prefix.star('.gemspec')),
-        )
-        lang_base.setup_cmd(
-            prefix,
-            (
-                'gem', 'install',
-                '--no-document', '--no-format-executable',
-                '--no-user-install',
-                '--install-dir', os.path.join(envdir, 'gems'),
-                '--bindir', os.path.join(envdir, 'gems', 'bin'),
-                *prefix.star('.gem'), *additional_dependencies,
-            ),
-        )
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix, hook.language_version):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/rust.py

@@ -1,113 +1,55 @@
-from __future__ import annotations
-
 import contextlib
-import functools
 import os.path
-import shutil
-import sys
-import tempfile
-import urllib.request
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Set
+from typing import Tuple
+
+import toml
 
 import pre_commit.constants as C
-from pre_commit import lang_base
-from pre_commit import parse_shebang
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output_b
-from pre_commit.util import make_executable
-from pre_commit.util import win_exe
 
 ENVIRONMENT_DIR = 'rustenv'
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 
 
-@functools.lru_cache(maxsize=1)
-def get_default_version() -> str:
-    # If rust is already installed, we can save a bunch of setup time by
-    # using the installed version.
-    #
-    # Just detecting the executable does not suffice, because if rustup is
-    # installed but no toolchain is available, then `cargo` exists but
-    # cannot be used without installing a toolchain first.
-    if cmd_output_b('cargo', '--version', check=False, cwd='/')[0] == 0:
-        return 'system'
-    else:
-        return C.DEFAULT
-
-
-def _rust_toolchain(language_version: str) -> str:
-    """Transform the language version into a rust toolchain version."""
-    if language_version == C.DEFAULT:
-        return 'stable'
-    else:
-        return language_version
-
-
-def get_env_patch(target_dir: str, version: str) -> PatchesT:
+def get_env_patch(target_dir: str) -> PatchesT:
     return (
         ('PATH', (os.path.join(target_dir, 'bin'), os.pathsep, Var('PATH'))),
-        # Only set RUSTUP_TOOLCHAIN if we don't want use the system's default
-        # toolchain
-        *(
-            (('RUSTUP_TOOLCHAIN', _rust_toolchain(version)),)
-            if version != 'system' else ()
-        ),
     )
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
-    with envcontext(get_env_patch(envdir, version)):
+def in_env(prefix: Prefix) -> Generator[None, None, None]:
+    target_dir = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT),
+    )
+    with envcontext(get_env_patch(target_dir)):
         yield
 
 
 def _add_dependencies(
-        prefix: Prefix,
-        additional_dependencies: set[str],
+        cargo_toml_path: str,
+        additional_dependencies: Set[str],
 ) -> None:
-    crates = []
-    for dep in additional_dependencies:
-        name, _, spec = dep.partition(':')
-        crate = f'{name}@{spec or "*"}'
-        crates.append(crate)
-
-    lang_base.setup_cmd(prefix, ('cargo', 'add', *crates))
-
-
-def install_rust_with_toolchain(toolchain: str, envdir: str) -> None:
-    with tempfile.TemporaryDirectory() as rustup_dir:
-        with envcontext((('CARGO_HOME', envdir), ('RUSTUP_HOME', rustup_dir))):
-            # acquire `rustup` if not present
-            if parse_shebang.find_executable('rustup') is None:
-                # We did not detect rustup and need to download it first.
-                if sys.platform == 'win32':  # pragma: win32 cover
-                    url = 'https://win.rustup.rs/x86_64'
-                else:  # pragma: win32 no cover
-                    url = 'https://sh.rustup.rs'
-
-                resp = urllib.request.urlopen(url)
-
-                rustup_init = os.path.join(rustup_dir, win_exe('rustup-init'))
-                with open(rustup_init, 'wb') as f:
-                    shutil.copyfileobj(resp, f)
-                make_executable(rustup_init)
-
-                # install rustup into `$CARGO_HOME/bin`
-                cmd_output_b(
-                    rustup_init, '-y', '--quiet', '--no-modify-path',
-                    '--default-toolchain', 'none',
-                )
-
-            cmd_output_b(
-                'rustup', 'toolchain', 'install', '--no-self-update',
-                toolchain,
-            )
+    with open(cargo_toml_path, 'r+') as f:
+        cargo_toml = toml.load(f)
+        cargo_toml.setdefault('dependencies', {})
+        for dep in additional_dependencies:
+            name, _, spec = dep.partition(':')
+            cargo_toml['dependencies'][name] = spec or '*'
+        f.seek(0)
+        toml.dump(cargo_toml, f)
+        f.truncate()
 
 
 def install_environment(
@@ -115,7 +57,10 @@ def install_environment(
         version: str,
         additional_dependencies: Sequence[str],
 ) -> None:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    helpers.assert_version_default('rust', version)
+    directory = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT),
+    )
 
     # There are two cases where we might want to specify more dependencies:
     # as dependencies for the library being built, and as binary packages
@@ -132,29 +77,30 @@ def install_environment(
     }
     lib_deps = set(additional_dependencies) - cli_deps
 
-    packages_to_install: set[tuple[str, ...]] = {('--path', '.')}
-    for cli_dep in cli_deps:
-        cli_dep = cli_dep.removeprefix('cli:')
-        package, _, crate_version = cli_dep.partition(':')
-        if crate_version != '':
-            packages_to_install.add((package, '--version', crate_version))
-        else:
-            packages_to_install.add((package,))
+    if len(lib_deps) > 0:
+        _add_dependencies(prefix.path('Cargo.toml'), lib_deps)
 
-    with contextlib.ExitStack() as ctx:
-        ctx.enter_context(in_env(prefix, version))
-
-        if version != 'system':
-            install_rust_with_toolchain(_rust_toolchain(version), envdir)
-
-            tmpdir = ctx.enter_context(tempfile.TemporaryDirectory())
-            ctx.enter_context(envcontext((('RUSTUP_HOME', tmpdir),)))
-
-        if len(lib_deps) > 0:
-            _add_dependencies(prefix, lib_deps)
+    with clean_path_on_failure(directory):
+        packages_to_install: Set[Tuple[str, ...]] = {('--path', '.')}
+        for cli_dep in cli_deps:
+            cli_dep = cli_dep[len('cli:'):]
+            package, _, version = cli_dep.partition(':')
+            if version != '':
+                packages_to_install.add((package, '--version', version))
+            else:
+                packages_to_install.add((package,))
 
         for args in packages_to_install:
             cmd_output_b(
-                'cargo', 'install', '--bins', '--root', envdir, *args,
+                'cargo', 'install', '--bins', '--root', directory, *args,
                 cwd=prefix.prefix_dir,
             )
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    with in_env(hook.prefix):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/script.py

@@ -0,0 +1,19 @@
+from typing import Sequence
+from typing import Tuple
+
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
+
+ENVIRONMENT_DIR = None
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
+install_environment = helpers.no_install
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    cmd = (hook.prefix.path(hook.cmd[0]), *hook.cmd[1:])
+    return helpers.run_xargs(hook, cmd, file_args, color=color)

pre_commit/languages/swift.py

@@ -1,25 +1,25 @@
-from __future__ import annotations
-
 import contextlib
 import os
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Generator
+from typing import Sequence
+from typing import Tuple
 
-from pre_commit import lang_base
+import pre_commit.constants as C
 from pre_commit.envcontext import envcontext
 from pre_commit.envcontext import PatchesT
 from pre_commit.envcontext import Var
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
 from pre_commit.prefix import Prefix
+from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output_b
 
+ENVIRONMENT_DIR = 'swift_env'
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
 BUILD_DIR = '.build'
 BUILD_CONFIG = 'release'
 
-ENVIRONMENT_DIR = 'swift_env'
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-run_hook = lang_base.basic_run_hook
-
 
 def get_env_patch(venv: str) -> PatchesT:  # pragma: win32 no cover
     bin_path = os.path.join(venv, BUILD_DIR, BUILD_CONFIG)
@@ -27,8 +27,10 @@ def get_env_patch(venv: str) -> PatchesT:  # pragma: win32 no cover
 
 
 @contextlib.contextmanager  # pragma: win32 no cover
-def in_env(prefix: Prefix, version: str) -> Generator[None]:
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+def in_env(prefix: Prefix) -> Generator[None, None, None]:
+    envdir = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT),
+    )
     with envcontext(get_env_patch(envdir)):
         yield
 
@@ -36,15 +38,27 @@ def in_env(prefix: Prefix, version: str) -> Generator[None]:
 def install_environment(
         prefix: Prefix, version: str, additional_dependencies: Sequence[str],
 ) -> None:  # pragma: win32 no cover
-    lang_base.assert_version_default('swift', version)
-    lang_base.assert_no_additional_deps('swift', additional_dependencies)
-    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+    helpers.assert_version_default('swift', version)
+    helpers.assert_no_additional_deps('swift', additional_dependencies)
+    directory = prefix.path(
+        helpers.environment_dir(ENVIRONMENT_DIR, C.DEFAULT),
+    )
 
     # Build the swift package
-    os.mkdir(envdir)
-    cmd_output_b(
-        'swift', 'build',
-        '--package-path', prefix.prefix_dir,
-        '-c', BUILD_CONFIG,
-        '--build-path', os.path.join(envdir, BUILD_DIR),
-    )
+    with clean_path_on_failure(directory):
+        os.mkdir(directory)
+        cmd_output_b(
+            'swift', 'build',
+            '-C', prefix.prefix_dir,
+            '-c', BUILD_CONFIG,
+            '--build-path', os.path.join(directory, BUILD_DIR),
+        )
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:  # pragma: win32 no cover
+    with in_env(hook.prefix):
+        return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/system.py

@@ -0,0 +1,19 @@
+from typing import Sequence
+from typing import Tuple
+
+from pre_commit.hook import Hook
+from pre_commit.languages import helpers
+
+
+ENVIRONMENT_DIR = None
+get_default_version = helpers.basic_get_default_version
+healthy = helpers.basic_healthy
+install_environment = helpers.no_install
+
+
+def run_hook(
+        hook: Hook,
+        file_args: Sequence[str],
+        color: bool,
+) -> Tuple[int, bytes]:
+    return helpers.run_xargs(hook, hook.cmd, file_args, color=color)

pre_commit/languages/unsupported.py

@@ -1,10 +0,0 @@
-from __future__ import annotations
-
-from pre_commit import lang_base
-
-ENVIRONMENT_DIR = None
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-install_environment = lang_base.no_install
-in_env = lang_base.no_env
-run_hook = lang_base.basic_run_hook

pre_commit/languages/unsupported_script.py

@@ -1,32 +0,0 @@
-from __future__ import annotations
-
-from collections.abc import Sequence
-
-from pre_commit import lang_base
-from pre_commit.prefix import Prefix
-
-ENVIRONMENT_DIR = None
-get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
-install_environment = lang_base.no_install
-in_env = lang_base.no_env
-
-
-def run_hook(
-        prefix: Prefix,
-        entry: str,
-        args: Sequence[str],
-        file_args: Sequence[str],
-        *,
-        is_local: bool,
-        require_serial: bool,
-        color: bool,
-) -> tuple[int, bytes]:
-    cmd = lang_base.hook_cmd(entry, args)
-    cmd = (prefix.path(cmd[0]), *cmd[1:])
-    return lang_base.run_xargs(
-        cmd,
-        file_args,
-        require_serial=require_serial,
-        color=color,
-    )

pre_commit/logging_handler.py

@@ -1,8 +1,6 @@
-from __future__ import annotations
-
 import contextlib
 import logging
-from collections.abc import Generator
+from typing import Generator
 
 from pre_commit import color
 from pre_commit import output
@@ -32,7 +30,7 @@ class LoggingHandler(logging.Handler):
 
 
 @contextlib.contextmanager
-def logging_handler(use_color: bool) -> Generator[None]:
+def logging_handler(use_color: bool) -> Generator[None, None, None]:
     handler = LoggingHandler(use_color)
     logger.addHandler(handler)
     logger.setLevel(logging.INFO)

pre_commit/main.py

@@ -1,16 +1,15 @@
-from __future__ import annotations
-
 import argparse
 import logging
 import os
 import sys
-from collections.abc import Sequence
+from typing import Any
+from typing import Optional
+from typing import Sequence
+from typing import Union
 
 import pre_commit.constants as C
-from pre_commit import clientlib
 from pre_commit import git
 from pre_commit.color import add_color_option
-from pre_commit.commands import hazmat
 from pre_commit.commands.autoupdate import autoupdate
 from pre_commit.commands.clean import clean
 from pre_commit.commands.gc import gc
@@ -23,8 +22,6 @@ from pre_commit.commands.migrate_config import migrate_config
 from pre_commit.commands.run import run
 from pre_commit.commands.sample_config import sample_config
 from pre_commit.commands.try_repo import try_repo
-from pre_commit.commands.validate_config import validate_config
-from pre_commit.commands.validate_manifest import validate_manifest
 from pre_commit.error_handler import error_handler
 from pre_commit.logging_handler import logging_handler
 from pre_commit.store import Store
@@ -38,13 +35,8 @@ logger = logging.getLogger('pre_commit')
 # pyvenv
 os.environ.pop('__PYVENV_LAUNCHER__', None)
 
-# https://github.com/getsentry/snuba/pull/5388
-os.environ.pop('PYTHONEXECUTABLE', None)
 
-COMMANDS_NO_GIT = {
-    'clean', 'gc', 'hazmat', 'init-templatedir', 'sample-config',
-    'validate-config', 'validate-manifest',
-}
+COMMANDS_NO_GIT = {'clean', 'gc', 'init-templatedir', 'sample-config'}
 
 
 def _add_config_option(parser: argparse.ArgumentParser) -> None:
@@ -54,19 +46,43 @@ def _add_config_option(parser: argparse.ArgumentParser) -> None:
     )
 
 
+class AppendReplaceDefault(argparse.Action):
+    def __init__(self, *args: Any, **kwargs: Any) -> None:
+        super().__init__(*args, **kwargs)
+        self.appended = False
+
+    def __call__(
+            self,
+            parser: argparse.ArgumentParser,
+            namespace: argparse.Namespace,
+            values: Union[str, Sequence[str], None],
+            option_string: Optional[str] = None,
+    ) -> None:
+        if not self.appended:
+            setattr(namespace, self.dest, [])
+            self.appended = True
+        getattr(namespace, self.dest).append(values)
+
+
 def _add_hook_type_option(parser: argparse.ArgumentParser) -> None:
     parser.add_argument(
-        '-t', '--hook-type',
-        choices=clientlib.HOOK_TYPES, action='append', dest='hook_types',
+        '-t', '--hook-type', choices=(
+            'pre-commit', 'pre-merge-commit', 'pre-push', 'prepare-commit-msg',
+            'commit-msg', 'post-commit', 'post-checkout', 'post-merge',
+            'post-rewrite',
+        ),
+        action=AppendReplaceDefault,
+        default=['pre-commit'],
+        dest='hook_types',
     )
 
 
 def _add_run_options(parser: argparse.ArgumentParser) -> None:
     parser.add_argument('hook', nargs='?', help='A single hook-id to run')
-    parser.add_argument('--verbose', '-v', action='store_true')
+    parser.add_argument('--verbose', '-v', action='store_true', default=False)
     mutex_group = parser.add_mutually_exclusive_group(required=False)
     mutex_group.add_argument(
-        '--all-files', '-a', action='store_true',
+        '--all-files', '-a', action='store_true', default=False,
         help='Run on all the files in the repo.',
     )
     mutex_group.add_argument(
@@ -78,14 +94,7 @@ def _add_run_options(parser: argparse.ArgumentParser) -> None:
         help='When hooks fail, run `git diff` directly afterward.',
     )
     parser.add_argument(
-        '--fail-fast', action='store_true',
-        help='Stop after the first failing hook.',
-    )
-    parser.add_argument(
-        '--hook-stage',
-        choices=clientlib.STAGES,
-        type=clientlib.transform_stage,
-        default='pre-commit',
+        '--hook-stage', choices=C.STAGES, default='commit',
         help='The stage during which the hook is fired.  One of %(choices)s',
     )
     parser.add_argument(
@@ -97,7 +106,7 @@ def _add_run_options(parser: argparse.ArgumentParser) -> None:
     parser.add_argument(
         '--from-ref', '--source', '-s',
         help=(
-            '(for usage with `--to-ref`) -- this option represents the '
+            '(for usage with `--from-ref`) -- this option represents the '
             'original ref in a `from_ref...to_ref` diff expression.  '
             'For `pre-push` hooks, this represents the branch you are pushing '
             'to.  '
@@ -108,42 +117,17 @@ def _add_run_options(parser: argparse.ArgumentParser) -> None:
     parser.add_argument(
         '--to-ref', '--origin', '-o',
         help=(
-            '(for usage with `--from-ref`) -- this option represents the '
+            '(for usage with `--to-ref`) -- this option represents the '
             'destination ref in a `from_ref...to_ref` diff expression.  '
             'For `pre-push` hooks, this represents the branch being pushed.  '
             'For `post-checkout` hooks, this represents the branch that is '
             'now checked out.'
         ),
     )
-    parser.add_argument(
-        '--pre-rebase-upstream', help=(
-            'The upstream from which the series was forked.'
-        ),
-    )
-    parser.add_argument(
-        '--pre-rebase-branch', help=(
-            'The branch being rebased, and is not set when  '
-            'rebasing the current branch.'
-        ),
-    )
     parser.add_argument(
         '--commit-msg-filename',
         help='Filename to check when running during `commit-msg`',
     )
-    parser.add_argument(
-        '--prepare-commit-message-source',
-        help=(
-            'Source of the commit message '
-            '(typically the second argument to .git/hooks/prepare-commit-msg)'
-        ),
-    )
-    parser.add_argument(
-        '--commit-object-name',
-        help=(
-            'Commit object name '
-            '(typically the third argument to .git/hooks/prepare-commit-msg)'
-        ),
-    )
     parser.add_argument(
         '--remote-name', help='Remote name used by `git push`.',
     )
@@ -178,10 +162,6 @@ def _adjust_args_and_chdir(args: argparse.Namespace) -> None:
         args.config = os.path.abspath(args.config)
     if args.command in {'run', 'try-repo'}:
         args.files = [os.path.abspath(filename) for filename in args.files]
-        if args.commit_msg_filename is not None:
-            args.commit_msg_filename = os.path.abspath(
-                args.commit_msg_filename,
-            )
     if args.command == 'try-repo' and os.path.exists(args.repo):
         args.repo = os.path.abspath(args.repo)
 
@@ -191,15 +171,11 @@ def _adjust_args_and_chdir(args: argparse.Namespace) -> None:
     args.config = os.path.relpath(args.config)
     if args.command in {'run', 'try-repo'}:
         args.files = [os.path.relpath(filename) for filename in args.files]
-        if args.commit_msg_filename is not None:
-            args.commit_msg_filename = os.path.relpath(
-                args.commit_msg_filename,
-            )
     if args.command == 'try-repo' and os.path.exists(args.repo):
         args.repo = os.path.relpath(args.repo)
 
 
-def main(argv: Sequence[str] | None = None) -> int:
+def main(argv: Optional[Sequence[str]] = None) -> int:
     argv = argv if argv is not None else sys.argv[1:]
     parser = argparse.ArgumentParser(prog='pre-commit')
 
@@ -212,20 +188,16 @@ def main(argv: Sequence[str] | None = None) -> int:
 
     subparsers = parser.add_subparsers(dest='command')
 
-    def _add_cmd(name: str, *, help: str) -> argparse.ArgumentParser:
-        parser = subparsers.add_parser(name, help=help)
-        add_color_option(parser)
-        return parser
-
-    autoupdate_parser = _add_cmd(
+    autoupdate_parser = subparsers.add_parser(
         'autoupdate',
         help="Auto-update pre-commit config to the latest repos' versions.",
     )
+    add_color_option(autoupdate_parser)
     _add_config_option(autoupdate_parser)
     autoupdate_parser.add_argument(
         '--bleeding-edge', action='store_true',
         help=(
-            'Update to the bleeding edge of `HEAD` instead of the latest '
+            'Update to the bleeding edge of `master` instead of the latest '
             'tagged version (the default behavior).'
         ),
     )
@@ -234,30 +206,38 @@ def main(argv: Sequence[str] | None = None) -> int:
         help='Store "frozen" hashes in `rev` instead of tag names',
     )
     autoupdate_parser.add_argument(
-        '--repo', dest='repos', action='append', metavar='REPO', default=[],
+        '--repo', dest='repos', action='append', metavar='REPO',
         help='Only update this repository -- may be specified multiple times.',
     )
-    autoupdate_parser.add_argument(
-        '-j', '--jobs', type=int, default=1,
-        help='Number of threads to use.  (default %(default)s).',
+
+    clean_parser = subparsers.add_parser(
+        'clean', help='Clean out pre-commit files.',
     )
+    add_color_option(clean_parser)
+    _add_config_option(clean_parser)
 
-    _add_cmd('clean', help='Clean out pre-commit files.')
-
-    _add_cmd('gc', help='Clean unused cached repos.')
-
-    hazmat_parser = _add_cmd(
-        'hazmat', help='Composable tools for rare use in hook `entry`.',
+    hook_impl_parser = subparsers.add_parser('hook-impl')
+    add_color_option(hook_impl_parser)
+    _add_config_option(hook_impl_parser)
+    hook_impl_parser.add_argument('--hook-type')
+    hook_impl_parser.add_argument('--hook-dir')
+    hook_impl_parser.add_argument(
+        '--skip-on-missing-config', action='store_true',
     )
-    hazmat.add_parsers(hazmat_parser)
+    hook_impl_parser.add_argument(dest='rest', nargs=argparse.REMAINDER)
 
-    init_templatedir_parser = _add_cmd(
+    gc_parser = subparsers.add_parser('gc', help='Clean unused cached repos.')
+    add_color_option(gc_parser)
+    _add_config_option(gc_parser)
+
+    init_templatedir_parser = subparsers.add_parser(
         'init-templatedir',
         help=(
             'Install hook script in a directory intended for use with '
             '`git config init.templateDir`.'
         ),
     )
+    add_color_option(init_templatedir_parser)
     _add_config_option(init_templatedir_parser)
     init_templatedir_parser.add_argument(
         'directory', help='The directory in which to write the hook script.',
@@ -270,7 +250,10 @@ def main(argv: Sequence[str] | None = None) -> int:
     )
     _add_hook_type_option(init_templatedir_parser)
 
-    install_parser = _add_cmd('install', help='Install the pre-commit script.')
+    install_parser = subparsers.add_parser(
+        'install', help='Install the pre-commit script.',
+    )
+    add_color_option(install_parser)
     _add_config_option(install_parser)
     install_parser.add_argument(
         '-f', '--overwrite', action='store_true',
@@ -285,14 +268,14 @@ def main(argv: Sequence[str] | None = None) -> int:
     )
     _add_hook_type_option(install_parser)
     install_parser.add_argument(
-        '--allow-missing-config', action='store_true',
+        '--allow-missing-config', action='store_true', default=False,
         help=(
             'Whether to allow a missing `pre-commit` configuration file '
             'or exit with a failure code.'
         ),
     )
 
-    install_hooks_parser = _add_cmd(
+    install_hooks_parser = subparsers.add_parser(
         'install-hooks',
         help=(
             'Install hook environments for all environments in the config '
@@ -300,24 +283,32 @@ def main(argv: Sequence[str] | None = None) -> int:
             'useful.'
         ),
     )
+    add_color_option(install_hooks_parser)
     _add_config_option(install_hooks_parser)
 
-    migrate_config_parser = _add_cmd(
+    migrate_config_parser = subparsers.add_parser(
         'migrate-config',
         help='Migrate list configuration to new map configuration.',
     )
+    add_color_option(migrate_config_parser)
     _add_config_option(migrate_config_parser)
 
-    run_parser = _add_cmd('run', help='Run hooks.')
+    run_parser = subparsers.add_parser('run', help='Run hooks.')
+    add_color_option(run_parser)
     _add_config_option(run_parser)
     _add_run_options(run_parser)
 
-    _add_cmd('sample-config', help=f'Produce a sample {C.CONFIG_FILE} file')
+    sample_config_parser = subparsers.add_parser(
+        'sample-config', help=f'Produce a sample {C.CONFIG_FILE} file',
+    )
+    add_color_option(sample_config_parser)
+    _add_config_option(sample_config_parser)
 
-    try_repo_parser = _add_cmd(
+    try_repo_parser = subparsers.add_parser(
         'try-repo',
         help='Try the hooks in a repository, useful for developing new hooks.',
     )
+    add_color_option(try_repo_parser)
     _add_config_option(try_repo_parser)
     try_repo_parser.add_argument(
         'repo', help='Repository to source hooks from.',
@@ -331,39 +322,18 @@ def main(argv: Sequence[str] | None = None) -> int:
     )
     _add_run_options(try_repo_parser)
 
-    uninstall_parser = _add_cmd(
+    uninstall_parser = subparsers.add_parser(
         'uninstall', help='Uninstall the pre-commit script.',
     )
+    add_color_option(uninstall_parser)
     _add_config_option(uninstall_parser)
     _add_hook_type_option(uninstall_parser)
 
-    validate_config_parser = _add_cmd(
-        'validate-config', help='Validate .pre-commit-config.yaml files',
-    )
-    validate_config_parser.add_argument('filenames', nargs='*')
-
-    validate_manifest_parser = _add_cmd(
-        'validate-manifest', help='Validate .pre-commit-hooks.yaml files',
-    )
-    validate_manifest_parser.add_argument('filenames', nargs='*')
-
-    # does not use `_add_cmd` because it doesn't use `--color`
     help = subparsers.add_parser(
         'help', help='Show help for a specific command.',
     )
     help.add_argument('help_cmd', nargs='?', help='Command to show help for.')
 
-    # not intended for users to call this directly
-    hook_impl_parser = subparsers.add_parser('hook-impl')
-    add_color_option(hook_impl_parser)
-    _add_config_option(hook_impl_parser)
-    hook_impl_parser.add_argument('--hook-type')
-    hook_impl_parser.add_argument('--hook-dir')
-    hook_impl_parser.add_argument(
-        '--skip-on-missing-config', action='store_true',
-    )
-    hook_impl_parser.add_argument(dest='rest', nargs=argparse.REMAINDER)
-
     # argparse doesn't really provide a way to use a `default` subparser
     if len(argv) == 0:
         argv = ['run']
@@ -377,26 +347,23 @@ def main(argv: Sequence[str] | None = None) -> int:
     with error_handler(), logging_handler(args.color):
         git.check_for_cygwin_mismatch()
 
-        store = Store()
-
         if args.command not in COMMANDS_NO_GIT:
             _adjust_args_and_chdir(args)
-            store.mark_config_used(args.config)
+
+        store = Store()
+        store.mark_config_used(args.config)
 
         if args.command == 'autoupdate':
             return autoupdate(
-                args.config,
+                args.config, store,
                 tags_only=not args.bleeding_edge,
                 freeze=args.freeze,
                 repos=args.repos,
-                jobs=args.jobs,
             )
         elif args.command == 'clean':
             return clean(store)
         elif args.command == 'gc':
             return gc(store)
-        elif args.command == 'hazmat':
-            return hazmat.impl(args)
         elif args.command == 'hook-impl':
             return hook_impl(
                 store,
@@ -432,14 +399,7 @@ def main(argv: Sequence[str] | None = None) -> int:
         elif args.command == 'try-repo':
             return try_repo(args)
         elif args.command == 'uninstall':
-            return uninstall(
-                config_file=args.config,
-                hook_types=args.hook_types,
-            )
-        elif args.command == 'validate-config':
-            return validate_config(args.filenames)
-        elif args.command == 'validate-manifest':
-            return validate_manifest(args.filenames)
+            return uninstall(hook_types=args.hook_types)
         else:
             raise NotImplementedError(
                 f'Command {args.command} not implemented.',
@@ -451,4 +411,4 @@ def main(argv: Sequence[str] | None = None) -> int:
 
 
 if __name__ == '__main__':
-    raise SystemExit(main())
+    exit(main())

pre_commit/meta_hooks/check_hooks_apply.py

@@ -1,7 +1,6 @@
-from __future__ import annotations
-
 import argparse
-from collections.abc import Sequence
+from typing import Optional
+from typing import Sequence
 
 import pre_commit.constants as C
 from pre_commit import git
@@ -21,14 +20,14 @@ def check_all_hooks_match_files(config_file: str) -> int:
     for hook in all_hooks(config, Store()):
         if hook.always_run or hook.language == 'fail':
             continue
-        elif not any(classifier.filenames_for_hook(hook)):
+        elif not classifier.filenames_for_hook(hook):
             print(f'{hook.id} does not apply to this repository')
             retv = 1
 
     return retv
 
 
-def main(argv: Sequence[str] | None = None) -> int:
+def main(argv: Optional[Sequence[str]] = None) -> int:
     parser = argparse.ArgumentParser()
     parser.add_argument('filenames', nargs='*', default=[C.CONFIG_FILE])
     args = parser.parse_args(argv)
@@ -40,4 +39,4 @@ def main(argv: Sequence[str] | None = None) -> int:
 
 
 if __name__ == '__main__':
-    raise SystemExit(main())
+    exit(main())

pre_commit/meta_hooks/check_useless_excludes.py

@@ -1,9 +1,7 @@
-from __future__ import annotations
-
 import argparse
 import re
-from collections.abc import Iterable
-from collections.abc import Sequence
+from typing import Optional
+from typing import Sequence
 
 from cfgv import apply_defaults
 
@@ -15,7 +13,7 @@ from pre_commit.commands.run import Classifier
 
 
 def exclude_matches_any(
-        filenames: Iterable[str],
+        filenames: Sequence[str],
         include: str,
         exclude: str,
 ) -> bool:
@@ -51,12 +49,11 @@ def check_useless_excludes(config_file: str) -> int:
             # Not actually a manifest dict, but this more accurately reflects
             # the defaults applied during runtime
             hook = apply_defaults(hook, MANIFEST_HOOK_DICT)
-            names = classifier.by_types(
-                classifier.filenames,
-                hook['types'],
-                hook['types_or'],
-                hook['exclude_types'],
-            )
+            names = classifier.filenames
+            types = hook['types']
+            types_or = hook['types_or']
+            exclude_types = hook['exclude_types']
+            names = classifier.by_types(names, types, types_or, exclude_types)
             include, exclude = hook['files'], hook['exclude']
             if not exclude_matches_any(names, include, exclude):
                 print(
@@ -68,7 +65,7 @@ def check_useless_excludes(config_file: str) -> int:
     return retv
 
 
-def main(argv: Sequence[str] | None = None) -> int:
+def main(argv: Optional[Sequence[str]] = None) -> int:
     parser = argparse.ArgumentParser()
     parser.add_argument('filenames', nargs='*', default=[C.CONFIG_FILE])
     args = parser.parse_args(argv)
@@ -80,4 +77,4 @@ def main(argv: Sequence[str] | None = None) -> int:
 
 
 if __name__ == '__main__':
-    raise SystemExit(main())
+    exit(main())

pre_commit/meta_hooks/identity.py

@@ -1,12 +1,11 @@
-from __future__ import annotations
-
 import sys
-from collections.abc import Sequence
+from typing import Optional
+from typing import Sequence
 
 from pre_commit import output
 
 
-def main(argv: Sequence[str] | None = None) -> int:
+def main(argv: Optional[Sequence[str]] = None) -> int:
     argv = argv if argv is not None else sys.argv[1:]
     for arg in argv:
         output.write_line(arg)
@@ -14,4 +13,4 @@ def main(argv: Sequence[str] | None = None) -> int:
 
 
 if __name__ == '__main__':
-    raise SystemExit(main())
+    exit(main())

pre_commit/output.py

@@ -1,9 +1,8 @@
-from __future__ import annotations
-
 import contextlib
 import sys
 from typing import Any
 from typing import IO
+from typing import Optional
 
 
 def write(s: str, stream: IO[bytes] = sys.stdout.buffer) -> None:
@@ -12,9 +11,9 @@ def write(s: str, stream: IO[bytes] = sys.stdout.buffer) -> None:
 
 
 def write_line_b(
-        s: bytes | None = None,
+        s: Optional[bytes] = None,
         stream: IO[bytes] = sys.stdout.buffer,
-        logfile_name: str | None = None,
+        logfile_name: Optional[str] = None,
 ) -> None:
     with contextlib.ExitStack() as exit_stack:
         output_streams = [stream]
@@ -29,5 +28,5 @@ def write_line_b(
             output_stream.flush()
 
 
-def write_line(s: str | None = None, **kwargs: Any) -> None:
+def write_line(s: Optional[str] = None, **kwargs: Any) -> None:
     write_line_b(s.encode() if s is not None else s, **kwargs)

pre_commit/parse_shebang.py

@@ -1,18 +1,21 @@
-from __future__ import annotations
-
 import os.path
-from collections.abc import Mapping
-from typing import NoReturn
+from typing import Mapping
+from typing import Optional
+from typing import Tuple
+from typing import TYPE_CHECKING
 
 from identify.identify import parse_shebang_from_file
 
+if TYPE_CHECKING:
+    from typing import NoReturn
+
 
 class ExecutableNotFoundError(OSError):
-    def to_output(self) -> tuple[int, bytes, None]:
+    def to_output(self) -> Tuple[int, bytes, None]:
         return (1, self.args[0].encode(), None)
 
 
-def parse_filename(filename: str) -> tuple[str, ...]:
+def parse_filename(filename: str) -> Tuple[str, ...]:
     if not os.path.exists(filename):
         return ()
     else:
@@ -20,13 +23,13 @@ def parse_filename(filename: str) -> tuple[str, ...]:
 
 
 def find_executable(
-        exe: str, *, env: Mapping[str, str] | None = None,
-) -> str | None:
+        exe: str, _environ: Optional[Mapping[str, str]] = None,
+) -> Optional[str]:
     exe = os.path.normpath(exe)
     if os.sep in exe:
         return exe
 
-    environ = env if env is not None else os.environ
+    environ = _environ if _environ is not None else os.environ
 
     if 'PATHEXT' in environ:
         exts = environ['PATHEXT'].split(os.pathsep)
@@ -43,12 +46,12 @@ def find_executable(
         return None
 
 
-def normexe(orig: str, *, env: Mapping[str, str] | None = None) -> str:
-    def _error(msg: str) -> NoReturn:
+def normexe(orig: str) -> str:
+    def _error(msg: str) -> 'NoReturn':
         raise ExecutableNotFoundError(f'Executable `{orig}` {msg}')
 
     if os.sep not in orig and (not os.altsep or os.altsep not in orig):
-        exe = find_executable(orig, env=env)
+        exe = find_executable(orig)
         if exe is None:
             _error('not found')
         return exe
@@ -62,11 +65,7 @@ def normexe(orig: str, *, env: Mapping[str, str] | None = None) -> str:
         return orig
 
 
-def normalize_cmd(
-        cmd: tuple[str, ...],
-        *,
-        env: Mapping[str, str] | None = None,
-) -> tuple[str, ...]:
+def normalize_cmd(cmd: Tuple[str, ...]) -> Tuple[str, ...]:
     """Fixes for the following issues on windows
     - https://bugs.python.org/issue8557
     - windows does not parse shebangs
@@ -74,12 +73,12 @@ def normalize_cmd(
     This function also makes deep-path shebangs work just fine
     """
     # Use PATH to determine the executable
-    exe = normexe(cmd[0], env=env)
+    exe = normexe(cmd[0])
 
     # Figure out the shebang from the resulting command
     cmd = parse_filename(exe) + (exe,) + cmd[1:]
 
     # This could have given us back another bare executable
-    exe = normexe(cmd[0], env=env)
+    exe = normexe(cmd[0])
 
     return (exe,) + cmd[1:]

pre_commit/prefix.py

@@ -1,7 +1,6 @@
-from __future__ import annotations
-
 import os.path
 from typing import NamedTuple
+from typing import Tuple
 
 
 class Prefix(NamedTuple):
@@ -13,6 +12,6 @@ class Prefix(NamedTuple):
     def exists(self, *parts: str) -> bool:
         return os.path.exists(self.path(*parts))
 
-    def star(self, end: str) -> tuple[str, ...]:
+    def star(self, end: str) -> Tuple[str, ...]:
         paths = os.listdir(self.prefix_dir)
         return tuple(path for path in paths if path.endswith(end))

pre_commit/repository.py

@@ -1,41 +1,40 @@
-from __future__ import annotations
-
 import json
 import logging
 import os
-from collections.abc import Sequence
 from typing import Any
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import Sequence
+from typing import Set
+from typing import Tuple
 
 import pre_commit.constants as C
-from pre_commit.all_languages import languages
 from pre_commit.clientlib import load_manifest
 from pre_commit.clientlib import LOCAL
 from pre_commit.clientlib import META
 from pre_commit.hook import Hook
-from pre_commit.lang_base import environment_dir
+from pre_commit.languages.all import languages
+from pre_commit.languages.helpers import environment_dir
 from pre_commit.prefix import Prefix
 from pre_commit.store import Store
-from pre_commit.util import clean_path_on_failure
+from pre_commit.util import parse_version
 from pre_commit.util import rmtree
 
 
 logger = logging.getLogger('pre_commit')
 
 
-def _state_filename_v1(venv: str) -> str:
-    return os.path.join(venv, '.install_state_v1')
-
-
-def _state_filename_v2(venv: str) -> str:
-    return os.path.join(venv, '.install_state_v2')
-
-
 def _state(additional_deps: Sequence[str]) -> object:
-    return {'additional_dependencies': additional_deps}
+    return {'additional_dependencies': sorted(additional_deps)}
 
 
-def _read_state(venv: str) -> object | None:
-    filename = _state_filename_v1(venv)
+def _state_filename(prefix: Prefix, venv: str) -> str:
+    return prefix.path(venv, f'.install_state_v{C.INSTALLED_STATE_VERSION}')
+
+
+def _read_state(prefix: Prefix, venv: str) -> Optional[object]:
+    filename = _state_filename(prefix, venv)
     if not os.path.exists(filename):
         return None
     else:
@@ -43,22 +42,26 @@ def _read_state(venv: str) -> object | None:
             return json.load(f)
 
 
+def _write_state(prefix: Prefix, venv: str, state: object) -> None:
+    state_filename = _state_filename(prefix, venv)
+    staging = f'{state_filename}staging'
+    with open(staging, 'w') as state_file:
+        state_file.write(json.dumps(state))
+    # Move the file into place atomically to indicate we've installed
+    os.replace(staging, state_filename)
+
+
 def _hook_installed(hook: Hook) -> bool:
     lang = languages[hook.language]
-    if lang.ENVIRONMENT_DIR is None:
-        return True
-
-    venv = environment_dir(
-        hook.prefix,
-        lang.ENVIRONMENT_DIR,
-        hook.language_version,
-    )
+    venv = environment_dir(lang.ENVIRONMENT_DIR, hook.language_version)
     return (
-        (
-            os.path.exists(_state_filename_v2(venv)) or
-            _read_state(venv) == _state(hook.additional_dependencies)
-        ) and
-        not lang.health_check(hook.prefix, hook.language_version)
+        venv is None or (
+            (
+                _read_state(hook.prefix, venv) ==
+                _state(hook.additional_dependencies)
+            ) and
+            lang.healthy(hook.prefix, hook.language_version)
+        )
     )
 
 
@@ -69,51 +72,43 @@ def _hook_install(hook: Hook) -> None:
 
     lang = languages[hook.language]
     assert lang.ENVIRONMENT_DIR is not None
-
-    venv = environment_dir(
-        hook.prefix,
-        lang.ENVIRONMENT_DIR,
-        hook.language_version,
-    )
+    venv = environment_dir(lang.ENVIRONMENT_DIR, hook.language_version)
 
     # There's potentially incomplete cleanup from previous runs
     # Clean it up!
-    if os.path.exists(venv):
-        rmtree(venv)
+    if hook.prefix.exists(venv):
+        rmtree(hook.prefix.path(venv))
 
-    with clean_path_on_failure(venv):
-        lang.install_environment(
-            hook.prefix, hook.language_version, hook.additional_dependencies,
+    lang.install_environment(
+        hook.prefix, hook.language_version, hook.additional_dependencies,
+    )
+    if not lang.healthy(hook.prefix, hook.language_version):
+        raise AssertionError(
+            f'BUG: expected environment for {hook.language} to be healthy() '
+            f'immediately after install, please open an issue describing '
+            f'your environment',
         )
-        health_error = lang.health_check(hook.prefix, hook.language_version)
-        if health_error:
-            raise AssertionError(
-                f'BUG: expected environment for {hook.language} to be healthy '
-                f'immediately after install, please open an issue describing '
-                f'your environment\n\n'
-                f'more info:\n\n{health_error}',
-            )
-
-        # TODO: remove v1 state writing, no longer needed after pre-commit 3.0
-        # Write our state to indicate we're installed
-        state_filename = _state_filename_v1(venv)
-        staging = f'{state_filename}staging'
-        with open(staging, 'w') as state_file:
-            state_file.write(json.dumps(_state(hook.additional_dependencies)))
-        # Move the file into place atomically to indicate we've installed
-        os.replace(staging, state_filename)
-
-        open(_state_filename_v2(venv), 'a+').close()
+    # Write our state to indicate we're installed
+    _write_state(hook.prefix, venv, _state(hook.additional_dependencies))
 
 
 def _hook(
-        *hook_dicts: dict[str, Any],
-        root_config: dict[str, Any],
-) -> dict[str, Any]:
+        *hook_dicts: Dict[str, Any],
+        root_config: Dict[str, Any],
+) -> Dict[str, Any]:
     ret, rest = dict(hook_dicts[0]), hook_dicts[1:]
     for dct in rest:
         ret.update(dct)
 
+    version = ret['minimum_pre_commit_version']
+    if parse_version(version) > parse_version(C.VERSION):
+        logger.error(
+            f'The hook `{ret["id"]}` requires pre-commit version {version} '
+            f'but version {C.VERSION} is installed.  '
+            f'Perhaps run `pip install --upgrade pre-commit`.',
+        )
+        exit(1)
+
     lang = ret['language']
     if ret['language_version'] == C.DEFAULT:
         ret['language_version'] = root_config['default_language_version'][lang]
@@ -145,10 +140,10 @@ def _hook(
 
 
 def _non_cloned_repository_hooks(
-        repo_config: dict[str, Any],
+        repo_config: Dict[str, Any],
         store: Store,
-        root_config: dict[str, Any],
-) -> tuple[Hook, ...]:
+        root_config: Dict[str, Any],
+) -> Tuple[Hook, ...]:
     def _prefix(language_name: str, deps: Sequence[str]) -> Prefix:
         language = languages[language_name]
         # pygrep / script / system / docker_image do not have
@@ -169,10 +164,10 @@ def _non_cloned_repository_hooks(
 
 
 def _cloned_repository_hooks(
-        repo_config: dict[str, Any],
+        repo_config: Dict[str, Any],
         store: Store,
-        root_config: dict[str, Any],
-) -> tuple[Hook, ...]:
+        root_config: Dict[str, Any],
+) -> Tuple[Hook, ...]:
     repo, rev = repo_config['repo'], repo_config['rev']
     manifest_path = os.path.join(store.clone(repo, rev), C.MANIFEST_FILE)
     by_id = {hook['id']: hook for hook in load_manifest(manifest_path)}
@@ -201,10 +196,10 @@ def _cloned_repository_hooks(
 
 
 def _repository_hooks(
-        repo_config: dict[str, Any],
+        repo_config: Dict[str, Any],
         store: Store,
-        root_config: dict[str, Any],
-) -> tuple[Hook, ...]:
+        root_config: Dict[str, Any],
+) -> Tuple[Hook, ...]:
     if repo_config['repo'] in {LOCAL, META}:
         return _non_cloned_repository_hooks(repo_config, store, root_config)
     else:
@@ -212,8 +207,8 @@ def _repository_hooks(
 
 
 def install_hook_envs(hooks: Sequence[Hook], store: Store) -> None:
-    def _need_installed() -> list[Hook]:
-        seen: set[tuple[Prefix, str, str, tuple[str, ...]]] = set()
+    def _need_installed() -> List[Hook]:
+        seen: Set[Tuple[Prefix, str, str, Tuple[str, ...]]] = set()
         ret = []
         for hook in hooks:
             if hook.install_key not in seen and not _hook_installed(hook):
@@ -229,7 +224,7 @@ def install_hook_envs(hooks: Sequence[Hook], store: Store) -> None:
             _hook_install(hook)
 
 
-def all_hooks(root_config: dict[str, Any], store: Store) -> tuple[Hook, ...]:
+def all_hooks(root_config: Dict[str, Any], store: Store) -> Tuple[Hook, ...]:
     return tuple(
         hook
         for repo in root_config['repos']

pre_commit/resources/empty_template_pre-commit-package-dev-1.rockspec

@@ -1,12 +0,0 @@
-package = "pre-commit-package"
-version = "dev-1"
-
-source = {
-   url = "git+ssh://git@github.com/pre-commit/pre-commit.git"
-}
-description = {}
-dependencies = {}
-build = {
-    type = "builtin",
-    modules = {},
-}

pre_commit/resources/empty_template_pubspec.yaml

@@ -1,4 +1,4 @@
 name: pre_commit_empty_pubspec
 environment:
-  sdk: '>=2.12.0'
+  sdk: '>=2.10.0'
 executables: {}

pre_commit/resources/empty_template_setup.py

@@ -1,4 +1,4 @@
 from setuptools import setup
 
 
-setup(name='pre-commit-placeholder-package', version='0.0.0', py_modules=[])
+setup(name='pre-commit-placeholder-package', version='0.0.0')

pre_commit/resources/hook-tmpl

@@ -1,20 +1,44 @@
-#!/usr/bin/env bash
+#!/usr/bin/env python3
 # File generated by pre-commit: https://pre-commit.com
 # ID: 138fd403232d2ddd5efb44317e38bf03
+import os
+import sys
+
+# we try our best, but the shebang of this script is difficult to determine:
+# - macos doesn't ship with python3
+# - windows executables are almost always `python.exe`
+# therefore we continue to support python2 for this small script
+if sys.version_info < (3, 3):
+    from distutils.spawn import find_executable as which
+else:
+    from shutil import which
+
+# work around https://github.com/Homebrew/homebrew-core/issues/30445
+os.environ.pop('__PYVENV_LAUNCHER__', None)
 
 # start templated
-INSTALL_PYTHON=''
-ARGS=(hook-impl)
+INSTALL_PYTHON = ''
+ARGS = ['hook-impl']
 # end templated
+ARGS.extend(('--hook-dir', os.path.realpath(os.path.dirname(__file__))))
+ARGS.append('--')
+ARGS.extend(sys.argv[1:])
 
-HERE="$(cd "$(dirname "$0")" && pwd)"
-ARGS+=(--hook-dir "$HERE" -- "$@")
+DNE = '`pre-commit` not found.  Did you forget to activate your virtualenv?'
+if os.access(INSTALL_PYTHON, os.X_OK):
+    CMD = [INSTALL_PYTHON, '-mpre_commit']
+elif which('pre-commit'):
+    CMD = ['pre-commit']
+else:
+    raise SystemExit(DNE)
 
-if [ -x "$INSTALL_PYTHON" ]; then
-    exec "$INSTALL_PYTHON" -mpre_commit "${ARGS[@]}"
-elif command -v pre-commit > /dev/null; then
-    exec pre-commit "${ARGS[@]}"
-else
-    echo '`pre-commit` not found.  Did you forget to activate your virtualenv?' 1>&2
-    exit 1
-fi
+CMD.extend(ARGS)
+if sys.platform == 'win32':  # https://bugs.python.org/issue19124
+    import subprocess
+
+    if sys.version_info < (3, 7):  # https://bugs.python.org/issue25942
+        raise SystemExit(subprocess.Popen(CMD).wait())
+    else:
+        raise SystemExit(subprocess.call(CMD))
+else:
+    os.execvp(CMD[0], CMD)

pre_commit/staged_files_only.py

@@ -1,13 +1,10 @@
-from __future__ import annotations
-
 import contextlib
 import logging
 import os.path
 import time
-from collections.abc import Generator
+from typing import Generator
 
 from pre_commit import git
-from pre_commit.errors import FatalError
 from pre_commit.util import CalledProcessError
 from pre_commit.util import cmd_output
 from pre_commit.util import cmd_output_b
@@ -16,12 +13,6 @@ from pre_commit.xargs import xargs
 
 logger = logging.getLogger('pre_commit')
 
-# without forcing submodule.recurse=0, changes in nested submodules will be
-# discarded if `submodule.recurse=1` is configured
-# we choose this instead of `--no-recurse-submodules` because it works on
-# versions of git before that option was added to `git checkout`
-_CHECKOUT_CMD = ('git', '-c', 'submodule.recurse=0', 'checkout', '--', '.')
-
 
 def _git_apply(patch: str) -> None:
     args = ('apply', '--whitespace=nowarn', patch)
@@ -33,7 +24,7 @@ def _git_apply(patch: str) -> None:
 
 
 @contextlib.contextmanager
-def _intent_to_add_cleared() -> Generator[None]:
+def _intent_to_add_cleared() -> Generator[None, None, None]:
     intent_to_add = git.intent_to_add_files()
     if intent_to_add:
         logger.warning('Unstaged intent-to-add files detected.')
@@ -48,23 +39,14 @@ def _intent_to_add_cleared() -> Generator[None]:
 
 
 @contextlib.contextmanager
-def _unstaged_changes_cleared(patch_dir: str) -> Generator[None]:
+def _unstaged_changes_cleared(patch_dir: str) -> Generator[None, None, None]:
     tree = cmd_output('git', 'write-tree')[1].strip()
-    diff_cmd = (
+    retcode, diff_stdout_binary, _ = cmd_output_b(
         'git', 'diff-index', '--ignore-submodules', '--binary',
         '--exit-code', '--no-color', '--no-ext-diff', tree, '--',
+        retcode=None,
     )
-    retcode, diff_stdout, diff_stderr = cmd_output_b(*diff_cmd, check=False)
-    if retcode == 0:
-        # There weren't any staged files so we don't need to do anything
-        # special
-        yield
-    elif retcode == 1 and not diff_stdout.strip():
-        # due to behaviour (probably a bug?) in git with crlf endings and
-        # autocrlf set to either `true` or `input` sometimes git will refuse
-        # to show a crlf-only diff to us :(
-        yield
-    elif retcode == 1 and diff_stdout.strip():
+    if retcode and diff_stdout_binary.strip():
         patch_filename = f'patch{int(time.time())}-{os.getpid()}'
         patch_filename = os.path.join(patch_dir, patch_filename)
         logger.warning('Unstaged files detected.')
@@ -72,13 +54,13 @@ def _unstaged_changes_cleared(patch_dir: str) -> Generator[None]:
         # Save the current unstaged changes as a patch
         os.makedirs(patch_dir, exist_ok=True)
         with open(patch_filename, 'wb') as patch_file:
-            patch_file.write(diff_stdout)
+            patch_file.write(diff_stdout_binary)
 
         # prevent recursive post-checkout hooks (#1418)
         no_checkout_env = dict(os.environ, _PRE_COMMIT_SKIP_POST_CHECKOUT='1')
+        cmd_output_b('git', 'checkout', '--', '.', env=no_checkout_env)
 
         try:
-            cmd_output_b(*_CHECKOUT_CMD, env=no_checkout_env)
             yield
         finally:
             # Try to apply the patch we saved
@@ -92,20 +74,18 @@ def _unstaged_changes_cleared(patch_dir: str) -> Generator[None]:
                 # We failed to apply the patch, presumably due to fixes made
                 # by hooks.
                 # Roll back the changes made by hooks.
-                cmd_output_b(*_CHECKOUT_CMD, env=no_checkout_env)
+                cmd_output_b('git', 'checkout', '--', '.', env=no_checkout_env)
                 _git_apply(patch_filename)
 
             logger.info(f'Restored changes from {patch_filename}.')
-    else:  # pragma: win32 no cover
-        # some error occurred while requesting the diff
-        e = CalledProcessError(retcode, diff_cmd, b'', diff_stderr)
-        raise FatalError(
-            f'pre-commit failed to diff -- perhaps due to permissions?\n\n{e}',
-        )
+    else:
+        # There weren't any staged files so we don't need to do anything
+        # special
+        yield
 
 
 @contextlib.contextmanager
-def staged_files_only(patch_dir: str) -> Generator[None]:
+def staged_files_only(patch_dir: str) -> Generator[None, None, None]:
     """Clear any unstaged changes from the git working directory inside this
     context.
     """

pre_commit/store.py

@@ -1,22 +1,23 @@
-from __future__ import annotations
-
 import contextlib
 import logging
 import os.path
 import sqlite3
 import tempfile
-from collections.abc import Callable
-from collections.abc import Generator
-from collections.abc import Sequence
+from typing import Callable
+from typing import Generator
+from typing import List
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
 
 import pre_commit.constants as C
-from pre_commit import clientlib
 from pre_commit import file_lock
 from pre_commit import git
 from pre_commit.util import CalledProcessError
 from pre_commit.util import clean_path_on_failure
 from pre_commit.util import cmd_output_b
 from pre_commit.util import resource_text
+from pre_commit.util import rmtree
 
 
 logger = logging.getLogger('pre_commit')
@@ -36,30 +37,10 @@ def _get_default_directory() -> str:
     return os.path.realpath(ret)
 
 
-_LOCAL_RESOURCES = (
-    'Cargo.toml', 'main.go', 'go.mod', 'main.rs', '.npmignore',
-    'package.json', 'pre-commit-package-dev-1.rockspec',
-    'pre_commit_placeholder_package.gemspec', 'setup.py',
-    'environment.yml', 'Makefile.PL', 'pubspec.yaml',
-    'renv.lock', 'renv/activate.R', 'renv/LICENSE.renv',
-)
-
-
-def _make_local_repo(directory: str) -> None:
-    for resource in _LOCAL_RESOURCES:
-        resource_dirname, resource_basename = os.path.split(resource)
-        contents = resource_text(f'empty_template_{resource_basename}')
-        target_dir = os.path.join(directory, resource_dirname)
-        target_file = os.path.join(target_dir, resource_basename)
-        os.makedirs(target_dir, exist_ok=True)
-        with open(target_file, 'w') as f:
-            f.write(contents)
-
-
 class Store:
     get_default_directory = staticmethod(_get_default_directory)
 
-    def __init__(self, directory: str | None = None) -> None:
+    def __init__(self, directory: Optional[str] = None) -> None:
         self.directory = directory or Store.get_default_directory()
         self.db_path = os.path.join(self.directory, 'db.db')
         self.readonly = (
@@ -95,13 +76,13 @@ class Store:
                     '    PRIMARY KEY (repo, ref)'
                     ');',
                 )
-                self._create_configs_table(db)
+                self._create_config_table(db)
 
             # Atomic file move
             os.replace(tmpfile, self.db_path)
 
     @contextlib.contextmanager
-    def exclusive_lock(self) -> Generator[None]:
+    def exclusive_lock(self) -> Generator[None, None, None]:
         def blocked_cb() -> None:  # pragma: no cover (tests are in-process)
             logger.info('Locking pre-commit directory')
 
@@ -111,8 +92,8 @@ class Store:
     @contextlib.contextmanager
     def connect(
             self,
-            db_path: str | None = None,
-    ) -> Generator[sqlite3.Connection]:
+            db_path: Optional[str] = None,
+    ) -> Generator[sqlite3.Connection, None, None]:
         db_path = db_path or self.db_path
         # sqlite doesn't close its fd with its contextmanager >.<
         # contextlib.closing fixes this.
@@ -125,7 +106,7 @@ class Store:
     @classmethod
     def db_repo_name(cls, repo: str, deps: Sequence[str]) -> str:
         if deps:
-            return f'{repo}:{",".join(deps)}'
+            return f'{repo}:{",".join(sorted(deps))}'
         else:
             return repo
 
@@ -136,10 +117,9 @@ class Store:
             deps: Sequence[str],
             make_strategy: Callable[[str], None],
     ) -> str:
-        original_repo = repo
         repo = self.db_repo_name(repo, deps)
 
-        def _get_result() -> str | None:
+        def _get_result() -> Optional[str]:
             # Check if we already exist
             with self.connect() as db:
                 result = db.execute(
@@ -169,9 +149,6 @@ class Store:
                     'INSERT INTO repos (repo, ref, path) VALUES (?, ?, ?)',
                     [repo, ref, directory],
                 )
-
-            clientlib.warn_for_stages_on_repo_init(original_repo, directory)
-
         return directory
 
     def _complete_clone(self, ref: str, git_cmd: Callable[..., None]) -> None:
@@ -209,12 +186,39 @@ class Store:
 
         return self._new_repo(repo, ref, deps, clone_strategy)
 
+    LOCAL_RESOURCES = (
+        'Cargo.toml', 'main.go', 'go.mod', 'main.rs', '.npmignore',
+        'package.json', 'pre_commit_placeholder_package.gemspec', 'setup.py',
+        'environment.yml', 'Makefile.PL', 'pubspec.yaml',
+        'renv.lock', 'renv/activate.R', 'renv/LICENSE.renv',
+    )
+
     def make_local(self, deps: Sequence[str]) -> str:
+        def make_local_strategy(directory: str) -> None:
+            for resource in self.LOCAL_RESOURCES:
+                resource_dirname, resource_basename = os.path.split(resource)
+                contents = resource_text(f'empty_template_{resource_basename}')
+                target_dir = os.path.join(directory, resource_dirname)
+                target_file = os.path.join(target_dir, resource_basename)
+                os.makedirs(target_dir, exist_ok=True)
+                with open(target_file, 'w') as f:
+                    f.write(contents)
+
+            env = git.no_git_env()
+
+            # initialize the git repository so it looks more like cloned repos
+            def _git_cmd(*args: str) -> None:
+                cmd_output_b('git', *args, cwd=directory, env=env)
+
+            git.init_repo(directory, '<<unknown>>')
+            _git_cmd('add', '.')
+            git.commit(repo=directory)
+
         return self._new_repo(
-            'local', C.LOCAL_REPO_VERSION, deps, _make_local_repo,
+            'local', C.LOCAL_REPO_VERSION, deps, make_local_strategy,
         )
 
-    def _create_configs_table(self, db: sqlite3.Connection) -> None:
+    def _create_config_table(self, db: sqlite3.Connection) -> None:
         db.executescript(
             'CREATE TABLE IF NOT EXISTS configs ('
             '   path TEXT NOT NULL,'
@@ -231,5 +235,28 @@ class Store:
             return
         with self.connect() as db:
             # TODO: eventually remove this and only create in _create
-            self._create_configs_table(db)
+            self._create_config_table(db)
             db.execute('INSERT OR IGNORE INTO configs VALUES (?)', (path,))
+
+    def select_all_configs(self) -> List[str]:
+        with self.connect() as db:
+            self._create_config_table(db)
+            rows = db.execute('SELECT path FROM configs').fetchall()
+            return [path for path, in rows]
+
+    def delete_configs(self, configs: List[str]) -> None:
+        with self.connect() as db:
+            rows = [(path,) for path in configs]
+            db.executemany('DELETE FROM configs WHERE path = ?', rows)
+
+    def select_all_repos(self) -> List[Tuple[str, str, str]]:
+        with self.connect() as db:
+            return db.execute('SELECT repo, ref, path from repos').fetchall()
+
+    def delete_repo(self, db_repo_name: str, ref: str, path: str) -> None:
+        with self.connect() as db:
+            db.execute(
+                'DELETE FROM repos WHERE repo = ? and ref = ?',
+                (db_repo_name, ref),
+            )
+        rmtree(path)

pre_commit/util.py

@@ -1,20 +1,45 @@
-from __future__ import annotations
-
 import contextlib
 import errno
-import importlib.resources
+import functools
 import os.path
 import shutil
 import stat
 import subprocess
 import sys
-from collections.abc import Callable
-from collections.abc import Generator
+import tempfile
 from types import TracebackType
 from typing import Any
+from typing import Callable
+from typing import Dict
+from typing import Generator
+from typing import IO
+from typing import Optional
+from typing import Tuple
+from typing import Type
+
+import yaml
 
 from pre_commit import parse_shebang
 
+if sys.version_info >= (3, 7):  # pragma: no cover (PY37+)
+    from importlib.resources import open_binary
+    from importlib.resources import read_text
+else:  # pragma: no cover (<PY37)
+    from importlib_resources import open_binary
+    from importlib_resources import read_text
+
+Loader = getattr(yaml, 'CSafeLoader', yaml.SafeLoader)
+yaml_load = functools.partial(yaml.load, Loader=Loader)
+Dumper = getattr(yaml, 'CSafeDumper', yaml.SafeDumper)
+
+
+def yaml_dump(o: Any, **kwargs: Any) -> str:
+    # when python/mypy#1484 is solved, this can be `functools.partial`
+    return yaml.dump(
+        o, Dumper=Dumper, default_flow_style=False, indent=4, sort_keys=False,
+        **kwargs,
+    )
+
 
 def force_bytes(exc: Any) -> bytes:
     with contextlib.suppress(TypeError):
@@ -25,7 +50,7 @@ def force_bytes(exc: Any) -> bytes:
 
 
 @contextlib.contextmanager
-def clean_path_on_failure(path: str) -> Generator[None]:
+def clean_path_on_failure(path: str) -> Generator[None, None, None]:
     """Cleans up the directory on an exceptional failure."""
     try:
         yield
@@ -35,9 +60,24 @@ def clean_path_on_failure(path: str) -> Generator[None]:
         raise
 
 
+@contextlib.contextmanager
+def tmpdir() -> Generator[str, None, None]:
+    """Contextmanager to create a temporary directory.  It will be cleaned up
+    afterwards.
+    """
+    tempdir = tempfile.mkdtemp()
+    try:
+        yield tempdir
+    finally:
+        rmtree(tempdir)
+
+
+def resource_bytesio(filename: str) -> IO[bytes]:
+    return open_binary('pre_commit.resources', filename)
+
+
 def resource_text(filename: str) -> str:
-    files = importlib.resources.files('pre_commit.resources')
-    return files.joinpath(filename).read_text()
+    return read_text('pre_commit.resources', filename)
 
 
 def make_executable(filename: str) -> None:
@@ -50,26 +90,29 @@ class CalledProcessError(RuntimeError):
     def __init__(
             self,
             returncode: int,
-            cmd: tuple[str, ...],
+            cmd: Tuple[str, ...],
+            expected_returncode: int,
             stdout: bytes,
-            stderr: bytes | None,
+            stderr: Optional[bytes],
     ) -> None:
-        super().__init__(returncode, cmd, stdout, stderr)
+        super().__init__(returncode, cmd, expected_returncode, stdout, stderr)
         self.returncode = returncode
         self.cmd = cmd
+        self.expected_returncode = expected_returncode
         self.stdout = stdout
         self.stderr = stderr
 
     def __bytes__(self) -> bytes:
-        def _indent_or_none(part: bytes | None) -> bytes:
+        def _indent_or_none(part: Optional[bytes]) -> bytes:
             if part:
-                return b'\n    ' + part.replace(b'\n', b'\n    ').rstrip()
+                return b'\n    ' + part.replace(b'\n', b'\n    ')
             else:
                 return b' (none)'
 
         return b''.join((
             f'command: {self.cmd!r}\n'.encode(),
             f'return code: {self.returncode}\n'.encode(),
+            f'expected return code: {self.expected_returncode}\n'.encode(),
             b'stdout:', _indent_or_none(self.stdout), b'\n',
             b'stderr:', _indent_or_none(self.stderr),
         ))
@@ -78,24 +121,24 @@ class CalledProcessError(RuntimeError):
         return self.__bytes__().decode()
 
 
-def _setdefault_kwargs(kwargs: dict[str, Any]) -> None:
+def _setdefault_kwargs(kwargs: Dict[str, Any]) -> None:
     for arg in ('stdin', 'stdout', 'stderr'):
         kwargs.setdefault(arg, subprocess.PIPE)
 
 
-def _oserror_to_output(e: OSError) -> tuple[int, bytes, None]:
+def _oserror_to_output(e: OSError) -> Tuple[int, bytes, None]:
     return 1, force_bytes(e).rstrip(b'\n') + b'\n', None
 
 
 def cmd_output_b(
         *cmd: str,
-        check: bool = True,
+        retcode: Optional[int] = 0,
         **kwargs: Any,
-) -> tuple[int, bytes, bytes | None]:
+) -> Tuple[int, bytes, Optional[bytes]]:
     _setdefault_kwargs(kwargs)
 
     try:
-        cmd = parse_shebang.normalize_cmd(cmd, env=kwargs.get('env'))
+        cmd = parse_shebang.normalize_cmd(cmd)
     except parse_shebang.ExecutableNotFoundError as e:
         returncode, stdout_b, stderr_b = e.to_output()
     else:
@@ -107,36 +150,36 @@ def cmd_output_b(
             stdout_b, stderr_b = proc.communicate()
             returncode = proc.returncode
 
-    if check and returncode:
-        raise CalledProcessError(returncode, cmd, stdout_b, stderr_b)
+    if retcode is not None and retcode != returncode:
+        raise CalledProcessError(returncode, cmd, retcode, stdout_b, stderr_b)
 
     return returncode, stdout_b, stderr_b
 
 
-def cmd_output(*cmd: str, **kwargs: Any) -> tuple[int, str, str | None]:
+def cmd_output(*cmd: str, **kwargs: Any) -> Tuple[int, str, Optional[str]]:
     returncode, stdout_b, stderr_b = cmd_output_b(*cmd, **kwargs)
     stdout = stdout_b.decode() if stdout_b is not None else None
     stderr = stderr_b.decode() if stderr_b is not None else None
     return returncode, stdout, stderr
 
 
-if sys.platform != 'win32':  # pragma: win32 no cover
+if os.name != 'nt':  # pragma: win32 no cover
     from os import openpty
     import termios
 
     class Pty:
         def __init__(self) -> None:
-            self.r: int | None = None
-            self.w: int | None = None
+            self.r: Optional[int] = None
+            self.w: Optional[int] = None
 
-        def __enter__(self) -> Pty:
+        def __enter__(self) -> 'Pty':
             self.r, self.w = openpty()
 
             # tty flags normally change \n to \r\n
-            attrs = termios.tcgetattr(self.w)
+            attrs = termios.tcgetattr(self.r)
             assert isinstance(attrs[1], int)
             attrs[1] &= ~(termios.ONLCR | termios.OPOST)
-            termios.tcsetattr(self.w, termios.TCSANOW, attrs)
+            termios.tcsetattr(self.r, termios.TCSANOW, attrs)
 
             return self
 
@@ -152,19 +195,19 @@ if sys.platform != 'win32':  # pragma: win32 no cover
 
         def __exit__(
                 self,
-                exc_type: type[BaseException] | None,
-                exc_value: BaseException | None,
-                traceback: TracebackType | None,
+                exc_type: Optional[Type[BaseException]],
+                exc_value: Optional[BaseException],
+                traceback: Optional[TracebackType],
         ) -> None:
             self.close_w()
             self.close_r()
 
     def cmd_output_p(
             *cmd: str,
-            check: bool = True,
+            retcode: Optional[int] = 0,
             **kwargs: Any,
-    ) -> tuple[int, bytes, bytes | None]:
-        assert check is False
+    ) -> Tuple[int, bytes, Optional[bytes]]:
+        assert retcode is None
         assert kwargs['stderr'] == subprocess.STDOUT, kwargs['stderr']
         _setdefault_kwargs(kwargs)
 
@@ -202,37 +245,29 @@ else:  # pragma: no cover
     cmd_output_p = cmd_output_b
 
 
-def _handle_readonly(
-        func: Callable[[str], object],
-        path: str,
-        exc: BaseException,
-) -> None:
-    if (
-            func in (os.rmdir, os.remove, os.unlink) and
-            isinstance(exc, OSError) and
-            exc.errno in {errno.EACCES, errno.EPERM}
-    ):
-        for p in (path, os.path.dirname(path)):
-            os.chmod(p, os.stat(p).st_mode | stat.S_IWUSR)
-        func(path)
-    else:
-        raise
-
-
-if sys.version_info < (3, 12):  # pragma: <3.12 cover
-    def _handle_readonly_old(
-        func: Callable[[str], object],
-        path: str,
-        excinfo: tuple[type[BaseException], BaseException, TracebackType],
+def rmtree(path: str) -> None:
+    """On windows, rmtree fails for readonly dirs."""
+    def handle_remove_readonly(
+            func: Callable[..., Any],
+            path: str,
+            exc: Tuple[Type[OSError], OSError, TracebackType],
     ) -> None:
-        return _handle_readonly(func, path, excinfo[1])
+        excvalue = exc[1]
+        if (
+                func in (os.rmdir, os.remove, os.unlink) and
+                excvalue.errno in {errno.EACCES, errno.EPERM}
+        ):
+            for p in (path, os.path.dirname(path)):
+                os.chmod(p, os.stat(p).st_mode | stat.S_IWUSR)
+            func(path)
+        else:
+            raise
+    shutil.rmtree(path, ignore_errors=False, onerror=handle_remove_readonly)
 
-    def rmtree(path: str) -> None:
-        shutil.rmtree(path, ignore_errors=False, onerror=_handle_readonly_old)
-else:  # pragma: >=3.12 cover
-    def rmtree(path: str) -> None:
-        """On windows, rmtree fails for readonly dirs."""
-        shutil.rmtree(path, ignore_errors=False, onexc=_handle_readonly)
+
+def parse_version(s: str) -> Tuple[int, ...]:
+    """poor man's version comparison"""
+    return tuple(int(p) for p in s.split('.'))
 
 
 def win_exe(s: str) -> str:

pre_commit/xargs.py

@@ -1,18 +1,18 @@
-from __future__ import annotations
-
 import concurrent.futures
 import contextlib
 import math
-import multiprocessing
 import os
 import subprocess
 import sys
-from collections.abc import Callable
-from collections.abc import Generator
-from collections.abc import Iterable
-from collections.abc import MutableMapping
-from collections.abc import Sequence
 from typing import Any
+from typing import Callable
+from typing import Generator
+from typing import Iterable
+from typing import List
+from typing import MutableMapping
+from typing import Optional
+from typing import Sequence
+from typing import Tuple
 from typing import TypeVar
 
 from pre_commit import parse_shebang
@@ -23,22 +23,7 @@ TArg = TypeVar('TArg')
 TRet = TypeVar('TRet')
 
 
-def cpu_count() -> int:
-    try:
-        # On systems that support it, this will return a more accurate count of
-        # usable CPUs for the current process, which will take into account
-        # cgroup limits
-        return len(os.sched_getaffinity(0))
-    except AttributeError:
-        pass
-
-    try:
-        return multiprocessing.cpu_count()
-    except NotImplementedError:
-        return 1
-
-
-def _environ_size(_env: MutableMapping[str, str] | None = None) -> int:
+def _environ_size(_env: Optional[MutableMapping[str, str]] = None) -> int:
     environ = _env if _env is not None else getattr(os, 'environb', os.environ)
     size = 8 * len(environ)  # number of pointers in `envp`
     for k, v in environ.items():
@@ -77,8 +62,8 @@ def partition(
         cmd: Sequence[str],
         varargs: Sequence[str],
         target_concurrency: int,
-        _max_length: int | None = None,
-) -> tuple[tuple[str, ...], ...]:
+        _max_length: Optional[int] = None,
+) -> Tuple[Tuple[str, ...], ...]:
     _max_length = _max_length or _get_platform_max_length()
 
     # Generally, we try to partition evenly into at least `target_concurrency`
@@ -88,7 +73,7 @@ def partition(
     cmd = tuple(cmd)
     ret = []
 
-    ret_cmd: list[str] = []
+    ret_cmd: List[str] = []
     # Reversed so arguments are in order
     varargs = list(reversed(varargs))
 
@@ -120,6 +105,7 @@ def partition(
 @contextlib.contextmanager
 def _thread_mapper(maxsize: int) -> Generator[
     Callable[[Callable[[TArg], TRet], Iterable[TArg]], Iterable[TRet]],
+    None, None,
 ]:
     if maxsize == 1:
         yield map
@@ -129,14 +115,14 @@ def _thread_mapper(maxsize: int) -> Generator[
 
 
 def xargs(
-        cmd: tuple[str, ...],
+        cmd: Tuple[str, ...],
         varargs: Sequence[str],
         *,
         color: bool = False,
         target_concurrency: int = 1,
         _max_length: int = _get_platform_max_length(),
         **kwargs: Any,
-) -> tuple[int, bytes]:
+) -> Tuple[int, bytes]:
     """A simplified implementation of xargs.
 
     color: Make a pty if on a platform that supports it
@@ -166,10 +152,10 @@ def xargs(
     partitions = partition(cmd, varargs, target_concurrency, _max_length)
 
     def run_cmd_partition(
-            run_cmd: tuple[str, ...],
-    ) -> tuple[int, bytes, bytes | None]:
+            run_cmd: Tuple[str, ...],
+    ) -> Tuple[int, bytes, Optional[bytes]]:
         return cmd_fn(
-            *run_cmd, check=False, stderr=subprocess.STDOUT, **kwargs,
+            *run_cmd, retcode=None, stderr=subprocess.STDOUT, **kwargs,
         )
 
     threads = min(len(partitions), target_concurrency)
@@ -177,8 +163,7 @@ def xargs(
         results = thread_map(run_cmd_partition, partitions)
 
         for proc_retcode, proc_out, _ in results:
-            if abs(proc_retcode) > abs(retcode):
-                retcode = proc_retcode
+            retcode = max(retcode, proc_retcode)
             stdout += proc_out
 
     return retcode, stdout

pre_commit/yaml.py

@@ -1,19 +0,0 @@
-from __future__ import annotations
-
-import functools
-from typing import Any
-
-import yaml
-
-Loader = getattr(yaml, 'CSafeLoader', yaml.SafeLoader)
-yaml_compose = functools.partial(yaml.compose, Loader=Loader)
-yaml_load = functools.partial(yaml.load, Loader=Loader)
-Dumper = getattr(yaml, 'CSafeDumper', yaml.SafeDumper)
-
-
-def yaml_dump(o: Any, **kwargs: Any) -> str:
-    # when python/mypy#1484 is solved, this can be `functools.partial`
-    return yaml.dump(
-        o, Dumper=Dumper, default_flow_style=False, indent=4, sort_keys=False,
-        **kwargs,
-    )

pre_commit/yaml_rewrite.py

@@ -1,52 +0,0 @@
-from __future__ import annotations
-
-from collections.abc import Generator
-from collections.abc import Iterable
-from typing import NamedTuple
-from typing import Protocol
-
-from yaml.nodes import MappingNode
-from yaml.nodes import Node
-from yaml.nodes import ScalarNode
-from yaml.nodes import SequenceNode
-
-
-class _Matcher(Protocol):
-    def match(self, n: Node) -> Generator[Node]: ...
-
-
-class MappingKey(NamedTuple):
-    k: str
-
-    def match(self, n: Node) -> Generator[Node]:
-        if isinstance(n, MappingNode):
-            for k, _ in n.value:
-                if k.value == self.k:
-                    yield k
-
-
-class MappingValue(NamedTuple):
-    k: str
-
-    def match(self, n: Node) -> Generator[Node]:
-        if isinstance(n, MappingNode):
-            for k, v in n.value:
-                if k.value == self.k:
-                    yield v
-
-
-class SequenceItem(NamedTuple):
-    def match(self, n: Node) -> Generator[Node]:
-        if isinstance(n, SequenceNode):
-            yield from n.value
-
-
-def _match(gen: Iterable[Node], m: _Matcher) -> Iterable[Node]:
-    return (n for src in gen for n in m.match(src))
-
-
-def match(n: Node, matcher: tuple[_Matcher, ...]) -> Generator[ScalarNode]:
-    gen: Iterable[Node] = (n,)
-    for m in matcher:
-        gen = _match(gen, m)
-    return (n for n in gen if isinstance(n, ScalarNode))

requirements-dev.txt

@@ -1,4 +1,4 @@
-covdefaults>=2.2
+covdefaults
 coverage
 distlib
 pytest

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = pre_commit
-version = 4.5.1
+version = 2.15.0
 description = A framework for managing and maintaining multi-language pre-commit hooks.
 long_description = file: README.md
 long_description_content_type = text/markdown
@@ -8,10 +8,15 @@ url = https://github.com/pre-commit/pre-commit
 author = Anthony Sottile
 author_email = asottile@umich.edu
 license = MIT
-license_files = LICENSE
+license_file = LICENSE
 classifiers =
+    License :: OSI Approved :: MIT License
     Programming Language :: Python :: 3
     Programming Language :: Python :: 3 :: Only
+    Programming Language :: Python :: 3.6
+    Programming Language :: Python :: 3.7
+    Programming Language :: Python :: 3.8
+    Programming Language :: Python :: 3.9
     Programming Language :: Python :: Implementation :: CPython
     Programming Language :: Python :: Implementation :: PyPy
 
@@ -22,8 +27,11 @@ install_requires =
     identify>=1.0.0
     nodeenv>=0.11.1
     pyyaml>=5.1
-    virtualenv>=20.10.0
-python_requires = >=3.10
+    toml
+    virtualenv>=20.0.8
+    importlib-metadata;python_version<"3.8"
+    importlib-resources;python_version<"3.7"
+python_requires = >=3.6.1
 
 [options.packages.find]
 exclude =
@@ -33,6 +41,8 @@ exclude =
 [options.entry_points]
 console_scripts =
     pre-commit = pre_commit.main:main
+    pre-commit-validate-config = pre_commit.clientlib:validate_config_main
+    pre-commit-validate-manifest = pre_commit.clientlib:validate_manifest_main
 
 [options.package_data]
 pre_commit.resources =
@@ -52,7 +62,7 @@ check_untyped_defs = true
 disallow_any_generics = true
 disallow_incomplete_defs = true
 disallow_untyped_defs = true
-enable_error_code = deprecated
+no_implicit_optional = true
 warn_redundant_casts = true
 warn_unused_ignores = true
 

setup.py

@@ -1,4 +1,2 @@
-from __future__ import annotations
-
 from setuptools import setup
 setup()

testing/auto_namedtuple.py

@@ -1,5 +1,3 @@
-from __future__ import annotations
-
 import collections
 
 

testing/fixtures.py

@@ -1,5 +1,3 @@
-from __future__ import annotations
-
 import contextlib
 import os.path
 import shutil
@@ -12,8 +10,8 @@ from pre_commit import git
 from pre_commit.clientlib import CONFIG_SCHEMA
 from pre_commit.clientlib import load_manifest
 from pre_commit.util import cmd_output
-from pre_commit.yaml import yaml_dump
-from pre_commit.yaml import yaml_load
+from pre_commit.util import yaml_dump
+from pre_commit.util import yaml_load
 from testing.util import get_resource_path
 from testing.util import git_commit
 
@@ -38,7 +36,7 @@ def copy_tree_to_path(src_dir, dest_dir):
 
 def git_dir(tempdir_factory):
     path = tempdir_factory.get()
-    cmd_output('git', '-c', 'init.defaultBranch=master', 'init', path)
+    cmd_output('git', 'init', path)
     return path
 
 

testing/gen-languages-all

@@ -0,0 +1,28 @@
+#!/usr/bin/env python3
+import sys
+
+LANGUAGES = [
+    'conda', 'coursier', 'dart', 'docker', 'docker_image', 'dotnet', 'fail',
+    'golang', 'node', 'perl', 'pygrep', 'python', 'r', 'ruby', 'rust',
+    'script', 'swift', 'system',
+]
+FIELDS = [
+    'ENVIRONMENT_DIR', 'get_default_version', 'healthy', 'install_environment',
+    'run_hook',
+]
+
+
+def main() -> int:
+    print(f'    # BEGIN GENERATED ({sys.argv[0]})')
+    for lang in LANGUAGES:
+        parts = [f'    {lang!r}: Language(name={lang!r}']
+        for k in FIELDS:
+            parts.append(f', {k}={lang}.{k}')
+        parts.append('),  # noqa: E501')
+        print(''.join(parts))
+    print('    # END GENERATED')
+    return 0
+
+
+if __name__ == '__main__':
+    exit(main())

testing/get-coursier.ps1

@@ -0,0 +1,11 @@
+$wc = New-Object System.Net.WebClient
+
+$coursier_url = "https://github.com/coursier/coursier/releases/download/v2.0.5/cs-x86_64-pc-win32.exe"
+$coursier_dest = "C:\coursier\cs.exe"
+$coursier_hash ="d63d497f7805261e1cd657b8aaa626f6b8f7264cdb68219b2e6be9dd882033a9"
+
+New-Item -Path "C:\" -Name "coursier" -ItemType "directory"
+$wc.DownloadFile($coursier_url, $coursier_dest)
+if ((Get-FileHash $coursier_dest -Algorithm SHA256).Hash -ne $coursier_hash) {
+    throw "Invalid coursier file"
+}

testing/get-coursier.sh

@@ -1,29 +1,15 @@
 #!/usr/bin/env bash
+# This is a script used in CI to install coursier
 set -euo pipefail
 
-if [ "$OSTYPE" = msys ]; then
-    URL='https://github.com/coursier/coursier/releases/download/v2.1.0-RC4/cs-x86_64-pc-win32.zip'
-    SHA256='0d07386ff0f337e3e6264f7dde29d137dda6eaa2385f29741435e0b93ccdb49d'
-    TARGET='/tmp/coursier/cs.zip'
-
-    unpack() {
-        unzip "$TARGET" -d /tmp/coursier
-        mv /tmp/coursier/cs-*.exe /tmp/coursier/cs.exe
-        cygpath -w /tmp/coursier >> "$GITHUB_PATH"
-    }
-else
-    URL='https://github.com/coursier/coursier/releases/download/v2.1.0-RC4/cs-x86_64-pc-linux.gz'
-    SHA256='176e92e08ab292531aa0c4993dbc9f2c99dec79578752f3b9285f54f306db572'
-    TARGET=/tmp/coursier/cs.gz
-
-    unpack() {
-        gunzip "$TARGET"
-        chmod +x /tmp/coursier/cs
-        echo /tmp/coursier >> "$GITHUB_PATH"
-    }
-fi
+COURSIER_URL="https://github.com/coursier/coursier/releases/download/v2.0.0/cs-x86_64-pc-linux"
+COURSIER_HASH="e2e838b75bc71b16bcb77ce951ad65660c89bda7957c79a0628ec7146d35122f"
+ARTIFACT="/tmp/coursier/cs"
 
 mkdir -p /tmp/coursier
-curl --location --silent --output "$TARGET" "$URL"
-echo "$SHA256 $TARGET" | sha256sum --check
-unpack
+rm -f "$ARTIFACT"
+curl --location --silent --output "$ARTIFACT" "$COURSIER_URL"
+echo "$COURSIER_HASH  $ARTIFACT" | sha256sum --check
+chmod ugo+x /tmp/coursier/cs
+
+echo '##vso[task.prependpath]/tmp/coursier'

testing/get-dart.sh

@@ -1,14 +1,14 @@
 #!/usr/bin/env bash
 set -euo pipefail
 
-VERSION=2.19.6
+VERSION=2.13.4
 
 if [ "$OSTYPE" = msys ]; then
     URL="https://storage.googleapis.com/dart-archive/channels/stable/release/${VERSION}/sdk/dartsdk-windows-x64-release.zip"
-    cygpath -w /tmp/dart-sdk/bin >> "$GITHUB_PATH"
+    echo "##vso[task.prependpath]$(cygpath -w /tmp/dart-sdk/bin)"
 else
     URL="https://storage.googleapis.com/dart-archive/channels/stable/release/${VERSION}/sdk/dartsdk-linux-x64-release.zip"
-    echo '/tmp/dart-sdk/bin' >> "$GITHUB_PATH"
+    echo '##vso[task.prependpath]/tmp/dart-sdk/bin'
 fi
 
 curl --silent --location --output /tmp/dart.zip "$URL"

testing/get-r.ps1

@@ -0,0 +1,6 @@
+$dir = $Env:Temp
+$urlR = "https://cran.r-project.org/bin/windows/base/old/4.0.4/R-4.0.4-win.exe"
+$outputR = "$dir\R-win.exe"
+$wcR = New-Object System.Net.WebClient
+$wcR.DownloadFile($urlR, $outputR)
+Start-Process -FilePath $outputR -ArgumentList "/S /v/qn"

testing/get-r.sh

@@ -0,0 +1,9 @@
+#!/usr/bin/env bash
+sudo apt install r-base
+# create empty folder for user library.
+# necessary for non-root users who have
+# never installed an R package before.
+# Alternatively, we require the renv
+# package to be installed already, then we can
+# omit that.
+Rscript -e 'dir.create(Sys.getenv("R_LIBS_USER"), recursive = TRUE)'

testing/get-swift.sh

@@ -0,0 +1,29 @@
+#!/usr/bin/env bash
+# This is a script used in CI to install swift
+set -euo pipefail
+
+. /etc/lsb-release
+if [ "$DISTRIB_CODENAME" = "bionic" ]; then
+    SWIFT_URL='https://swift.org/builds/swift-5.1.3-release/ubuntu1804/swift-5.1.3-RELEASE/swift-5.1.3-RELEASE-ubuntu18.04.tar.gz'
+    SWIFT_HASH='ac82ccd773fe3d586fc340814e31e120da1ff695c6a712f6634e9cc720769610'
+else
+    echo "unknown dist: ${DISTRIB_CODENAME}" 1>&2
+    exit 1
+fi
+
+check() {
+    echo "$SWIFT_HASH  $TGZ" | sha256sum --check
+}
+
+TGZ="$HOME/.swift/swift.tar.gz"
+mkdir -p "$(dirname "$TGZ")"
+if ! check >& /dev/null; then
+    rm -f "$TGZ"
+    curl --location --silent --output "$TGZ" "$SWIFT_URL"
+    check
+fi
+
+mkdir -p /tmp/swift
+tar -xf "$TGZ" --strip 1 --directory /tmp/swift
+
+echo '##vso[task.prependpath]/tmp/swift/usr/bin'

testing/language_helpers.py

@@ -1,40 +0,0 @@
-from __future__ import annotations
-
-import os
-from collections.abc import Sequence
-
-from pre_commit.lang_base import Language
-from pre_commit.prefix import Prefix
-
-
-def run_language(
-        path: os.PathLike[str],
-        language: Language,
-        exe: str,
-        args: Sequence[str] = (),
-        file_args: Sequence[str] = (),
-        version: str | None = None,
-        deps: Sequence[str] = (),
-        is_local: bool = False,
-        require_serial: bool = True,
-        color: bool = False,
-) -> tuple[int, bytes]:
-    prefix = Prefix(str(path))
-    version = version or language.get_default_version()
-
-    if language.ENVIRONMENT_DIR is not None:
-        language.install_environment(prefix, version, deps)
-        health_error = language.health_check(prefix, version)
-        assert health_error is None, health_error
-    with language.in_env(prefix, version):
-        ret, out = language.run_hook(
-            prefix,
-            exe,
-            args,
-            file_args,
-            is_local=is_local,
-            require_serial=require_serial,
-            color=color,
-        )
-        out = out.replace(b'\r\n', b'\n')
-        return ret, out

testing/languages

@@ -1,92 +0,0 @@
-#!/usr/bin/env python3
-from __future__ import annotations
-
-import argparse
-import concurrent.futures
-import json
-import os.path
-import subprocess
-import sys
-
-EXCLUDED = frozenset((
-    ('windows-latest', 'docker'),
-    ('windows-latest', 'docker_image'),
-    ('windows-latest', 'lua'),
-    ('windows-latest', 'swift'),
-))
-
-
-def _always_run() -> frozenset[str]:
-    ret = ['.github/workflows/languages.yaml', 'testing/languages']
-    ret.extend(
-        os.path.join('pre_commit/resources', fname)
-        for fname in os.listdir('pre_commit/resources')
-    )
-    return frozenset(ret)
-
-
-def _lang_files(lang: str) -> frozenset[str]:
-    prog = f'''\
-import json
-import os.path
-import sys
-
-import pre_commit.languages.{lang}
-import tests.languages.{lang}_test
-
-modules = sorted(
-    os.path.relpath(v.__file__)
-    for k, v in sys.modules.items()
-    if k.startswith(('pre_commit.', 'tests.', 'testing.'))
-)
-print(json.dumps(modules))
-'''
-    out = json.loads(subprocess.check_output((sys.executable, '-c', prog)))
-    return frozenset(out)
-
-
-def main() -> int:
-    parser = argparse.ArgumentParser()
-    parser.add_argument('--all', action='store_true')
-    args = parser.parse_args()
-
-    langs = [
-        os.path.splitext(fname)[0]
-        for fname in sorted(os.listdir('pre_commit/languages'))
-        if fname.endswith('.py') and fname != '__init__.py'
-    ]
-
-    triggers_all = _always_run()
-    for fname in triggers_all:
-        assert os.path.exists(fname), fname
-
-    if not args.all:
-        with concurrent.futures.ThreadPoolExecutor(os.cpu_count()) as exe:
-            by_lang = {
-                lang: files | triggers_all
-                for lang, files in zip(langs, exe.map(_lang_files, langs))
-            }
-
-        diff_cmd = ('git', 'diff', '--name-only', 'origin/main...HEAD')
-        files = set(subprocess.check_output(diff_cmd).decode().splitlines())
-
-        langs = [
-            lang
-            for lang, lang_files in by_lang.items()
-            if lang_files & files
-        ]
-
-    matched = [
-        {'os': os, 'language': lang}
-        for os in ('windows-latest', 'ubuntu-latest')
-        for lang in langs
-        if (os, lang) not in EXCLUDED
-    ]
-
-    with open(os.environ['GITHUB_OUTPUT'], 'a') as f:
-        f.write(f'languages={json.dumps(matched)}\n')
-    return 0
-
-
-if __name__ == '__main__':
-    raise SystemExit(main())