Merge pull request #3599 from pre-commit/pre-commit-ci-update-config

[pre-commit.ci] pre-commit autoupdate

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

@@ -6,4 +6,4 @@ runs:
   using: composite
   steps:
   - uses: asottile/workflows/.github/actions/latest-git@v1.4.0
-    if: inputs.env == 'py38' && runner.os == 'Linux'
+    if: inputs.env == 'py39' && runner.os == 'Linux'

.github/workflows/languages.yaml

@@ -3,7 +3,7 @@ name: languages
 on:
   push:
     branches: [main, test-me-*]
-    tags:
+    tags: '*'
   pull_request:
 
 concurrency:
@@ -21,7 +21,7 @@ jobs:
         fetch-depth: 0
     - uses: actions/setup-python@v4
       with:
-        python-version: 3.8
+        python-version: '3.10'
     - name: install deps
       run: python -mpip install -e . -r requirements-dev.txt
     - name: vars
@@ -36,10 +36,10 @@ jobs:
       matrix:
         include: ${{ fromJSON(needs.vars.outputs.languages) }}
     steps:
-    - uses: asottile/workflows/.github/actions/fast-checkout@v1.4.0
+    - uses: asottile/workflows/.github/actions/fast-checkout@v1.8.1
     - uses: actions/setup-python@v4
       with:
-        python-version: 3.8
+        python-version: '3.10'
 
     - run: echo "$CONDA\Scripts" >> "$GITHUB_PATH"
       shell: bash
@@ -63,8 +63,10 @@ jobs:
         echo 'C:\Strawberry\c\bin' >> "$GITHUB_PATH"
       shell: bash
       if: matrix.os == 'windows-latest' && matrix.language == 'perl'
-    - run: testing/get-swift.sh
-      if: matrix.os == 'ubuntu-latest' && matrix.language == 'swift'
+    - 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

.github/workflows/main.yml

@@ -3,7 +3,7 @@ name: main
 on:
   push:
     branches: [main, test-me-*]
-    tags:
+    tags: '*'
   pull_request:
 
 concurrency:
@@ -12,12 +12,12 @@ concurrency:
 
 jobs:
   main-windows:
-    uses: asottile/workflows/.github/workflows/tox.yml@v1.4.0
+    uses: asottile/workflows/.github/workflows/tox.yml@v1.8.1
     with:
-      env: '["py38"]'
+      env: '["py310"]'
       os: windows-latest
   main-linux:
-    uses: asottile/workflows/.github/workflows/tox.yml@v1.4.0
+    uses: asottile/workflows/.github/workflows/tox.yml@v1.8.1
     with:
-      env: '["py38", "py39", "py310"]'
+      env: '["py310", "py311", "py312", "py313"]'
       os: ubuntu-latest

.pre-commit-config.yaml

@@ -1,6 +1,6 @@
 repos:
 -   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.4.0
+    rev: v6.0.0
     hooks:
     -   id: trailing-whitespace
     -   id: end-of-file-fixer
@@ -10,36 +10,35 @@ repos:
     -   id: name-tests-test
     -   id: requirements-txt-fixer
 -   repo: https://github.com/asottile/setup-cfg-fmt
-    rev: v2.2.0
+    rev: v3.2.0
     hooks:
     -   id: setup-cfg-fmt
--   repo: https://github.com/asottile/reorder_python_imports
-    rev: v3.9.0
+-   repo: https://github.com/asottile/reorder-python-imports
+    rev: v3.16.0
     hooks:
     -   id: reorder-python-imports
-        exclude: ^(pre_commit/resources/|testing/resources/python3_hooks_repo/)
-        args: [--py38-plus, --add-import, 'from __future__ import annotations']
+        exclude: ^pre_commit/resources/
+        args: [--py310-plus, --add-import, 'from __future__ import annotations']
 -   repo: https://github.com/asottile/add-trailing-comma
-    rev: v2.4.0
+    rev: v4.0.0
     hooks:
     -   id: add-trailing-comma
-        args: [--py36-plus]
 -   repo: https://github.com/asottile/pyupgrade
-    rev: v3.3.2
+    rev: v3.21.2
     hooks:
     -   id: pyupgrade
-        args: [--py38-plus]
--   repo: https://github.com/pre-commit/mirrors-autopep8
-    rev: v2.0.2
+        args: [--py310-plus]
+-   repo: https://github.com/hhatto/autopep8
+    rev: v2.3.2
     hooks:
     -   id: autopep8
 -   repo: https://github.com/PyCQA/flake8
-    rev: 6.0.0
+    rev: 7.3.0
     hooks:
     -   id: flake8
 -   repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v1.2.0
+    rev: v1.19.1
     hooks:
     -   id: mypy
-        additional_dependencies: [types-all]
+        additional_dependencies: [types-pyyaml]
         exclude: ^testing/resources/

CHANGELOG.md

@@ -1,3 +1,246 @@
+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
 ==================
 

CONTRIBUTING.md

@@ -92,7 +92,7 @@ language, for example:
 
 here are the apis that should be implemented for a language
 
-Note that these are also documented in [`pre_commit/languages/all.py`](https://github.com/pre-commit/pre-commit/blob/main/pre_commit/languages/all.py)
+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)
 
 #### `ENVIRONMENT_DIR`
 
@@ -111,7 +111,7 @@ 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 = helpers.basic_default_version
+get_default_version = lang_base.basic_default_version
 ```
 
 `python` is currently the only language which implements this api
@@ -125,7 +125,7 @@ healthy.
 You generally don't need to implement this on a first pass and can just use:
 
 ```python
-health_check = helpers.basic_healthy_check
+health_check = lang_base.basic_health_check
 ```
 
 `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 = helpers.no_install`
+- (0th / 3rd class): `install_environment = lang_base.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`

pre_commit/all_languages.py

@@ -9,6 +9,8 @@ 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
@@ -17,9 +19,9 @@ 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.languages import unsupported
+from pre_commit.languages import unsupported_script
 
 
 languages: dict[str, Language] = {
@@ -31,6 +33,8 @@ languages: dict[str, Language] = {
     'dotnet': dotnet,
     'fail': fail,
     'golang': golang,
+    'haskell': haskell,
+    'julia': julia,
     'lua': lua,
     'node': node,
     'perl': perl,
@@ -39,10 +43,8 @@ languages: dict[str, Language] = {
     'r': r,
     'ruby': ruby,
     'rust': rust,
-    'script': script,
     'swift': swift,
-    'system': system,
-    # TODO: fully deprecate `python_venv`
-    'python_venv': python,
+    'unsupported': unsupported,
+    'unsupported_script': unsupported_script,
 }
 language_names = sorted(languages)

pre_commit/clientlib.py

@@ -2,12 +2,14 @@ from __future__ import annotations
 
 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 Sequence
 
 import cfgv
 from identify.identify import ALL_TAGS
@@ -70,6 +72,43 @@ 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]
@@ -78,11 +117,12 @@ class StagesMigrationNoDefault(NamedTuple):
         if self.key not in dct:
             return
 
-        val = dct[self.key]
-        cfgv.check_array(cfgv.check_any)(val)
+        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)
+            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:
@@ -99,13 +139,108 @@ class StagesMigration(StagesMigrationNoDefault):
         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)
+
+
 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),
-    cfgv.Required('language', cfgv.check_one_of(language_names)),
+    LanguageMigrationRequired('language', cfgv.check_one_of(language_names)),
     cfgv.Optional('alias', cfgv.check_string, ''),
 
     cfgv.Optional('files', check_string_regex, ''),
@@ -124,7 +259,6 @@ MANIFEST_HOOK_DICT = cfgv.Map(
     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('verbose', cfgv.check_bool, False),
@@ -136,10 +270,19 @@ 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=yaml_load,
+    load_strategy=_load_manifest_forward_compat,
     exc_tp=InvalidManifestError,
 )
 
@@ -261,12 +404,20 @@ class NotAllowed(cfgv.OptionalNoDefault):
             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 system
-    cfgv.Optional('language', cfgv.check_one_of({'system'}), 'system'),
+    # language must be `unsupported`
+    cfgv.Optional(
+        'language', cfgv.check_one_of({'unsupported'}), 'unsupported',
+    ),
     # entry cannot be overridden
     NotAllowed('entry', cfgv.check_any),
     *(
@@ -283,6 +434,7 @@ META_HOOK_DICT = cfgv.Map(
         item
         for item in MANIFEST_HOOK_DICT.items
     ),
+    *_COMMON_HOOK_WARNINGS,
 )
 CONFIG_HOOK_DICT = cfgv.Map(
     'Hook', 'id',
@@ -298,18 +450,17 @@ CONFIG_HOOK_DICT = cfgv.Map(
         for item in MANIFEST_HOOK_DICT.items
         if item.key != 'id'
         if item.key != 'stages'
+        if item.key != 'language'  # remove
     ),
     StagesMigrationNoDefault('stages', []),
-    OptionalSensibleRegexAtHook('files', cfgv.check_string),
-    OptionalSensibleRegexAtHook('exclude', cfgv.check_string),
+    LanguageMigration('language', cfgv.check_one_of(language_names)),  # remove
+    *_COMMON_HOOK_WARNINGS,
 )
 LOCAL_HOOK_DICT = cfgv.Map(
     'Hook', 'id',
 
     *MANIFEST_HOOK_DICT.items,
-
-    OptionalSensibleRegexAtHook('files', cfgv.check_string),
-    OptionalSensibleRegexAtHook('exclude', cfgv.check_string),
+    *_COMMON_HOOK_WARNINGS,
 )
 CONFIG_REPO_DICT = cfgv.Map(
     'Repository', 'repo',
@@ -345,6 +496,13 @@ DEFAULT_LANGUAGE_VERSION = cfgv.Map(
 CONFIG_SCHEMA = cfgv.Map(
     'Config', None,
 
+    # 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.RequiredRecurse('repos', cfgv.Array(CONFIG_REPO_DICT)),
     cfgv.Optional(
         'default_install_hook_types',
@@ -355,14 +513,10 @@ CONFIG_SCHEMA = cfgv.Map(
         '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.Optional(
-        'minimum_pre_commit_version',
-        cfgv.check_and(cfgv.check_string, check_min_version),
-        '0',
-    ),
     cfgv.WarnAdditionalKeys(
         (
             'repos',

pre_commit/commands/autoupdate.py

@@ -4,9 +4,9 @@ import concurrent.futures
 import os.path
 import re
 import tempfile
+from collections.abc import Sequence
 from typing import Any
 from typing import NamedTuple
-from typing import Sequence
 
 import pre_commit.constants as C
 from pre_commit import git
@@ -67,6 +67,8 @@ class RevInfo(NamedTuple):
                     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

pre_commit/commands/gc.py

@@ -12,6 +12,7 @@ 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(
@@ -26,7 +27,8 @@ 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'])
@@ -56,34 +58,41 @@ def _mark_used_repos(
             ))
 
 
-def _gc_repos(store: Store) -> int:
-    configs = store.select_all_configs()
-    repos = store.select_all_repos()
+def _gc(store: Store) -> int:
+    with store.exclusive_lock(), store.connect() as db:
+        store._create_configs_table(db)
 
-    # 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)]
+        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)
 
-    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)
+        configs_rows = db.execute('SELECT path FROM configs').fetchall()
+        configs = [path for path, in configs_rows]
 
-    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)
+        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)
 
 
 def gc(store: Store) -> int:
-    with store.exclusive_lock():
-        repos_removed = _gc_repos(store)
-    output.write_line(f'{repos_removed} repo(s) removed.')
+    output.write_line(f'{_gc(store)} repo(s) removed.')
     return 0

pre_commit/commands/hazmat.py

@@ -0,0 +1,95 @@
+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

@@ -4,7 +4,7 @@ import argparse
 import os.path
 import subprocess
 import sys
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit.commands.run import run
 from pre_commit.envcontext import envcontext
@@ -106,6 +106,7 @@ def _ns(
         hook=None,
         verbose=False,
         show_diff_on_failure=False,
+        fail_fast=False,
     )
 
 

pre_commit/commands/install_uninstall.py

@@ -103,8 +103,7 @@ def _install_hook_script(
 
         hook_file.write(before + TEMPLATE_START)
         hook_file.write(f'INSTALL_PYTHON={shlex.quote(sys.executable)}\n')
-        # TODO: python3.8+: shlex.join
-        args_s = ' '.join(shlex.quote(part) for part in args)
+        args_s = shlex.join(args)
         hook_file.write(f'ARGS=({args_s})\n')
         hook_file.write(TEMPLATE_END + after)
     make_executable(hook_path)

pre_commit/commands/migrate_config.py

@@ -1,13 +1,21 @@
 from __future__ import annotations
 
-import re
+import functools
+import itertools
 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
 
 
 def _is_header_line(line: str) -> bool:
@@ -38,16 +46,69 @@ def _migrate_map(contents: str) -> str:
     return contents
 
 
-def _migrate_sha_to_rev(contents: str) -> str:
-    return re.sub(r'(\n\s+)sha:', r'\1rev:', contents)
+def _preserve_style(n: ScalarNode, *, s: str) -> str:
+    style = n.style or ''
+    return f'{style}{s}{style}'
 
 
-def _migrate_python_venv(contents: str) -> str:
-    return re.sub(
-        r'(\n\s+)language: python_venv\b',
-        r'\1language: python',
-        contents,
+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_config(config_file: str, quiet: bool = False) -> int:
@@ -62,8 +123,7 @@ def migrate_config(config_file: str, quiet: bool = False) -> int:
                 raise cfgv.ValidationError(str(e))
 
     contents = _migrate_map(contents)
-    contents = _migrate_sha_to_rev(contents)
-    contents = _migrate_python_venv(contents)
+    contents = _migrate_composed(contents)
 
     if contents != orig_contents:
         with open(config_file, 'w') as f:

pre_commit/commands/run.py

@@ -9,10 +9,11 @@ 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 MutableMapping
-from typing import Sequence
 
 from identify.identify import tags_from_path
 
@@ -57,37 +58,36 @@ def _full_msg(
 
 
 def filter_by_include_exclude(
-        names: Collection[str],
+        names: Iterable[str],
         include: str,
         exclude: str,
-) -> list[str]:
+) -> Generator[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: Collection[str]) -> None:
+    def __init__(self, filenames: Iterable[str]) -> None:
         self.filenames = [f for f in filenames if os.path.lexists(f)]
 
-    @functools.lru_cache(maxsize=None)
+    @functools.cache
     def _types_for_file(self, filename: str) -> set[str]:
         return tags_from_path(filename)
 
     def by_types(
             self,
-            names: Sequence[str],
-            types: Collection[str],
-            types_or: Collection[str],
-            exclude_types: Collection[str],
-    ) -> list[str]:
+            names: Iterable[str],
+            types: Iterable[str],
+            types_or: Iterable[str],
+            exclude_types: Iterable[str],
+    ) -> Generator[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 (
@@ -95,24 +95,24 @@ class Classifier:
                     (not types_or or tags & types_or) and
                     not tags & exclude_types
             ):
-                ret.append(filename)
-        return ret
+                yield filename
 
-    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,
+    def filenames_for_hook(self, hook: Hook) -> Generator[str]:
+        return self.by_types(
+            filter_by_include_exclude(
+                self.filenames,
+                hook.files,
+                hook.exclude,
+            ),
             hook.types,
             hook.types_or,
             hook.exclude_types,
         )
-        return tuple(names)
 
     @classmethod
     def from_config(
             cls,
-            filenames: Collection[str],
+            filenames: Iterable[str],
             include: str,
             exclude: str,
     ) -> Classifier:
@@ -121,7 +121,7 @@ class Classifier:
         # 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)
 
@@ -148,7 +148,7 @@ def _run_single_hook(
         verbose: bool,
         use_color: bool,
 ) -> tuple[bool, bytes]:
-    filenames = classifier.filenames_for_hook(hook)
+    filenames = tuple(classifier.filenames_for_hook(hook))
 
     if hook.id in skips or hook.alias in skips:
         output.write(
@@ -187,7 +187,7 @@ def _run_single_hook(
 
         if not hook.pass_filenames:
             filenames = ()
-        time_before = time.time()
+        time_before = time.monotonic()
         language = languages[hook.language]
         with language.in_env(hook.prefix, hook.language_version):
             retcode, out = language.run_hook(
@@ -199,7 +199,7 @@ def _run_single_hook(
                 require_serial=hook.require_serial,
                 color=use_color,
             )
-        duration = round(time.time() - time_before, 2) or 0
+        duration = round(time.monotonic() - time_before, 2) or 0
         diff_after = _get_diff()
 
         # if the hook makes changes, fail the commit
@@ -250,7 +250,7 @@ def _compute_cols(hooks: Sequence[Hook]) -> int:
     return max(cols, 80)
 
 
-def _all_filenames(args: argparse.Namespace) -> Collection[str]:
+def _all_filenames(args: argparse.Namespace) -> Iterable[str]:
     # these hooks do not operate on files
     if args.hook_stage in {
         'post-checkout', 'post-commit', 'post-merge', 'post-rewrite',
@@ -298,7 +298,8 @@ def _run_hooks(
             verbose=args.verbose, use_color=args.color,
         )
         retval |= current_retval
-        if retval and (config['fail_fast'] or hook.fail_fast):
+        fail_fast = (config['fail_fast'] or hook.fail_fast or args.fail_fast)
+        if current_retval and fail_fast:
             break
     if retval and args.show_diff_on_failure and prior_diff:
         if args.all_files:

pre_commit/commands/validate_config.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit import clientlib
 

pre_commit/commands/validate_manifest.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit import clientlib
 

pre_commit/envcontext.py

@@ -3,10 +3,9 @@ from __future__ import annotations
 import contextlib
 import enum
 import os
-from typing import Generator
-from typing import MutableMapping
+from collections.abc import Generator
+from collections.abc import MutableMapping
 from typing import NamedTuple
-from typing import Tuple
 from typing import Union
 
 _Unset = enum.Enum('_Unset', 'UNSET')
@@ -18,9 +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:
@@ -34,7 +33,7 @@ def format_env(parts: SubstitutionT, env: MutableMapping[str, str]) -> str:
 def envcontext(
         patch: PatchesT,
         _env: MutableMapping[str, str] | None = None,
-) -> Generator[None, None, None]:
+) -> Generator[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

@@ -5,7 +5,7 @@ import functools
 import os.path
 import sys
 import traceback
-from typing import Generator
+from collections.abc import Generator
 from typing import IO
 
 import pre_commit.constants as C
@@ -68,7 +68,7 @@ def _log_and_exit(
 
 
 @contextlib.contextmanager
-def error_handler() -> Generator[None, None, None]:
+def error_handler() -> Generator[None]:
     try:
         yield
     except (Exception, KeyboardInterrupt) as e:

pre_commit/file_lock.py

@@ -3,8 +3,8 @@ from __future__ import annotations
 import contextlib
 import errno
 import sys
-from typing import Callable
-from typing import Generator
+from collections.abc import Callable
+from collections.abc import Generator
 
 
 if sys.platform == 'win32':  # pragma: no cover (windows)
@@ -20,7 +20,7 @@ if sys.platform == 'win32':  # pragma: no cover (windows)
     def _locked(
             fileno: int,
             blocked_cb: Callable[[], None],
-    ) -> Generator[None, None, None]:
+    ) -> Generator[None]:
         try:
             msvcrt.locking(fileno, msvcrt.LK_NBLCK, _region)
         except OSError:
@@ -53,7 +53,7 @@ else:  # pragma: win32 no cover
     def _locked(
             fileno: int,
             blocked_cb: Callable[[], None],
-    ) -> Generator[None, None, None]:
+    ) -> Generator[None]:
         try:
             fcntl.flock(fileno, fcntl.LOCK_EX | fcntl.LOCK_NB)
         except OSError:  # pragma: no cover (tests are single-threaded)
@@ -69,7 +69,7 @@ else:  # pragma: win32 no cover
 def lock(
         path: str,
         blocked_cb: Callable[[], None],
-) -> Generator[None, None, None]:
+) -> Generator[None]:
     with open(path, 'a+') as f:
         with _locked(f.fileno(), blocked_cb):
             yield

pre_commit/git.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 import logging
 import os.path
 import sys
-from typing import Mapping
+from collections.abc import Mapping
 
 from pre_commit.errors import FatalError
 from pre_commit.util import CalledProcessError
@@ -126,7 +126,7 @@ 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)
@@ -219,7 +219,7 @@ def check_for_cygwin_mismatch() -> None:
 
         if is_cygwin_python ^ is_cygwin_git:
             exe_type = {True: '(cygwin)', False: '(windows)'}
-            logger.warn(
+            logger.warning(
                 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'

pre_commit/hook.py

@@ -1,9 +1,9 @@
 from __future__ import annotations
 
 import logging
+from collections.abc import Sequence
 from typing import Any
 from typing import NamedTuple
-from typing import Sequence
 
 from pre_commit.prefix import Prefix
 

pre_commit/lang_base.py

@@ -5,12 +5,13 @@ 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 Generator
 from typing import NoReturn
 from typing import Protocol
-from typing import Sequence
 
 import pre_commit.constants as C
 from pre_commit import parse_shebang
@@ -127,7 +128,7 @@ def no_install(
 
 
 @contextlib.contextmanager
-def no_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+def no_env(prefix: Prefix, version: str) -> Generator[None]:
     yield
 
 
@@ -171,7 +172,10 @@ def run_xargs(
 
 
 def hook_cmd(entry: str, args: Sequence[str]) -> tuple[str, ...]:
-    return (*shlex.split(entry), *args)
+    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(

pre_commit/languages/conda.py

@@ -3,8 +3,8 @@ from __future__ import annotations
 import contextlib
 import os
 import sys
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
@@ -41,7 +41,7 @@ def get_env_patch(env: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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

pre_commit/languages/coursier.py

@@ -2,8 +2,8 @@ from __future__ import annotations
 
 import contextlib
 import os.path
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
@@ -70,7 +70,7 @@ def get_env_patch(target_dir: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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

pre_commit/languages/dart.py

@@ -4,8 +4,8 @@ import contextlib
 import os.path
 import shutil
 import tempfile
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
@@ -29,7 +29,7 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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

pre_commit/languages/docker.py

@@ -1,9 +1,12 @@
 from __future__ import annotations
 
+import contextlib
+import functools
 import hashlib
 import json
 import os
-from typing import Sequence
+import re
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.prefix import Prefix
@@ -16,31 +19,33 @@ 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
 
-
-def _is_in_docker() -> bool:
-    try:
-        with open('/proc/1/cgroup', 'rb') as f:
-            return b'docker' in f.read()
-    except FileNotFoundError:
-        return False
+_HOSTNAME_MOUNT_RE = re.compile(
+    rb"""
+    /containers
+    (?:/overlay-containers)?
+    /([a-z0-9]{64})
+    (?:/userdata)?
+    /hostname
+    """,
+    re.VERBOSE,
+)
 
 
-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_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()
+
+    return None
 
 
 def _get_docker_path(path: str) -> str:
-    if not _is_in_docker():
-        return path
-
     container_id = _get_container_id()
+    if container_id is None:
+        return path
 
     try:
         _, out, _ = cmd_output_b('docker', 'inspect', container_id)
@@ -101,17 +106,47 @@ def install_environment(
     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 ()
+
     try:
         return ('-u', f'{os.getuid()}:{os.getgid()}')
     except AttributeError:
         return ()
 
 
-def docker_cmd() -> tuple[str, ...]:  # pragma: win32 no cover
+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
     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
@@ -139,7 +174,7 @@ def run_hook(
 
     entry_tag = ('--entrypoint', entry_exe, docker_tag(prefix))
     return lang_base.run_xargs(
-        (*docker_cmd(), *entry_tag, *cmd_rest),
+        (*docker_cmd(color=color), *entry_tag, *cmd_rest),
         file_args,
         require_serial=require_serial,
         color=color,

pre_commit/languages/docker_image.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.languages.docker import docker_cmd
@@ -23,7 +23,7 @@ def run_hook(
         require_serial: bool,
         color: bool,
 ) -> tuple[int, bytes]:  # pragma: win32 no cover
-    cmd = docker_cmd() + lang_base.hook_cmd(entry, args)
+    cmd = docker_cmd(color=color) + lang_base.hook_cmd(entry, args)
     return lang_base.run_xargs(
         cmd,
         file_args,

pre_commit/languages/dotnet.py

@@ -6,8 +6,8 @@ import re
 import tempfile
 import xml.etree.ElementTree
 import zipfile
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
@@ -30,14 +30,14 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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
 
 
 @contextlib.contextmanager
-def _nuget_config_no_sources() -> Generator[str, None, None]:
+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:

pre_commit/languages/fail.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.prefix import Prefix

pre_commit/languages/golang.py

@@ -12,17 +12,18 @@ 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 Generator
 from typing import IO
 from typing import Protocol
-from typing import Sequence
 
 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 Var
+from pre_commit.git import no_git_env
 from pre_commit.prefix import Prefix
 from pre_commit.util import cmd_output
 from pre_commit.util import rmtree
@@ -74,6 +75,7 @@ def get_env_patch(venv: str, version: str) -> PatchesT:
 
     return (
         ('GOROOT', os.path.join(venv, '.go')),
+        ('GOTOOLCHAIN', 'local'),
         (
             'PATH', (
                 os.path.join(venv, 'bin'), os.pathsep,
@@ -88,8 +90,7 @@ def _infer_go_version(version: str) -> str:
     if version != C.DEFAULT:
         return version
     resp = urllib.request.urlopen('https://go.dev/dl/?mode=json')
-    # TODO: 3.9+ .removeprefix('go')
-    return json.load(resp)[0]['version'][2:]
+    return json.load(resp)[0]['version'].removeprefix('go')
 
 
 def _get_url(version: str) -> str:
@@ -120,7 +121,7 @@ def _install_go(version: str, dest: str) -> None:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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
@@ -141,9 +142,10 @@ def install_environment(
     else:
         gopath = env_dir
 
-    env = dict(os.environ, GOPATH=gopath)
+    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'],

pre_commit/languages/haskell.py

@@ -0,0 +1,56 @@
+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/julia.py

@@ -0,0 +1,133 @@
+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

@@ -3,8 +3,8 @@ from __future__ import annotations
 import contextlib
 import os
 import sys
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
@@ -44,7 +44,7 @@ def get_env_patch(d: str) -> PatchesT:  # pragma: win32 no cover
 
 
 @contextlib.contextmanager  # pragma: win32 no cover
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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

pre_commit/languages/node.py

@@ -4,8 +4,8 @@ import contextlib
 import functools
 import os
 import sys
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 import pre_commit.constants as C
 from pre_commit import lang_base
@@ -59,7 +59,7 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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
@@ -93,7 +93,7 @@ def install_environment(
         # install as if we installed from git
 
         local_install_cmd = (
-            'npm', 'install', '--dev', '--prod',
+            'npm', 'install', '--include=dev', '--include=prod',
             '--ignore-prepublish', '--no-progress', '--no-save',
         )
         lang_base.setup_cmd(prefix, local_install_cmd)

pre_commit/languages/perl.py

@@ -3,8 +3,8 @@ from __future__ import annotations
 import contextlib
 import os
 import shlex
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
@@ -33,7 +33,7 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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

pre_commit/languages/pygrep.py

@@ -3,9 +3,9 @@ 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 Pattern
-from typing import Sequence
 
 from pre_commit import lang_base
 from pre_commit import output

pre_commit/languages/python.py

@@ -4,8 +4,8 @@ import contextlib
 import functools
 import os
 import sys
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 import pre_commit.constants as C
 from pre_commit import lang_base
@@ -24,7 +24,7 @@ ENVIRONMENT_DIR = 'py_env'
 run_hook = lang_base.basic_run_hook
 
 
-@functools.lru_cache(maxsize=None)
+@functools.cache
 def _version_info(exe: str) -> str:
     prog = 'import sys;print(".".join(str(p) for p in sys.version_info))'
     try:
@@ -65,7 +65,7 @@ def _find_by_py_launcher(
         version: str,
 ) -> str | None:  # pragma: no cover (windows only)
     if version.startswith('python'):
-        num = version[len('python'):]
+        num = version.removeprefix('python')
         cmd = ('py', f'-{num}', '-c', 'import sys; print(sys.executable)')
         env = dict(os.environ, PYTHONIOENCODING='UTF-8')
         try:
@@ -75,6 +75,13 @@ 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:
         _, exe = os.path.split(path.lower())
@@ -100,18 +107,25 @@ def _find_by_sys_executable() -> str | None:
 
 @functools.lru_cache(maxsize=1)
 def get_default_version() -> str:  # pragma: no cover (platform dependent)
-    # First attempt from `sys.executable` (or the realpath)
-    exe = _find_by_sys_executable()
-    if exe:
-        return exe
+    v_major = f'{sys.version_info[0]}'
+    v_minor = f'{sys.version_info[0]}.{sys.version_info[1]}'
 
-    # Next try the `pythonX.X` executable
-    exe = f'python{sys.version_info[0]}.{sys.version_info[1]}'
-    if find_executable(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
 
-    if _find_by_py_launcher(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
 
     # We tried!
     return C.DEFAULT
@@ -124,7 +138,7 @@ def _sys_executable_matches(version: str) -> bool:
         return False
 
     try:
-        info = tuple(int(p) for p in version[len('python'):].split('.'))
+        info = tuple(int(p) for p in version.removeprefix('python').split('.'))
     except ValueError:
         return False
 
@@ -152,7 +166,7 @@ def norm_version(version: str) -> str | None:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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

pre_commit/languages/r.py

@@ -4,21 +4,120 @@ import contextlib
 import os
 import shlex
 import shutil
-from typing import Generator
-from typing import Sequence
+import tempfile
+import textwrap
+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
+from pre_commit.util import cmd_output
 from pre_commit.util import win_exe
 
 ENVIRONMENT_DIR = 'renv'
-RSCRIPT_OPTS = ('--no-save', '--no-restore', '--no-site-file', '--no-environ')
 get_default_version = lang_base.basic_get_default_version
-health_check = lang_base.basic_health_check
+
+_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
 
 
 def get_env_patch(venv: str) -> PatchesT:
@@ -29,7 +128,7 @@ def get_env_patch(venv: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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
@@ -85,7 +184,7 @@ def _cmd_from_hook(
     _entry_validate(cmd)
 
     cmd_part = _prefix_if_file_entry(cmd, prefix, is_local=is_local)
-    return (cmd[0], *RSCRIPT_OPTS, *cmd_part, *args)
+    return (cmd[0], *_RENV_ACTIVATED_OPTS, *cmd_part, *args)
 
 
 def install_environment(
@@ -93,6 +192,8 @@ 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)
@@ -126,21 +227,19 @@ def install_environment(
             renv::install(prefix_dir)
         }}
         """
-
-    cmd_output_b(
-        _rscript_exec(), '--vanilla', '-e',
-        _inline_r_setup(r_code_inst_environment),
-        cwd=env_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))'
-        with in_env(prefix, version):
-            cmd_output_b(
-                _rscript_exec(), *RSCRIPT_OPTS, '-e',
-                _inline_r_setup(r_code_inst_add),
-                *additional_dependencies,
-                cwd=env_dir,
-            )
+        _execute_r_in_renv(
+            code=r_code_inst_add, prefix=prefix, version=version,
+            args=additional_dependencies,
+            cwd=env_dir,
+        )
 
 
 def _inline_r_setup(code: str) -> str:
@@ -148,11 +247,16 @@ 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 = f"""\
-    options(install.packages.compile.from.source = "never", pkgType = "binary")
-    {code}
-    """
-    return with_option
+    with_option = [
+        textwrap.dedent("""\
+        options(
+            install.packages.compile.from.source = "never",
+            pkgType = "binary"
+        )
+        """),
+        code,
+    ]
+    return '\n'.join(with_option)
 
 
 def run_hook(

pre_commit/languages/ruby.py

@@ -6,9 +6,9 @@ import importlib.resources
 import os.path
 import shutil
 import tarfile
-from typing import Generator
+from collections.abc import Generator
+from collections.abc import Sequence
 from typing import IO
-from typing import Sequence
 
 import pre_commit.constants as C
 from pre_commit import lang_base
@@ -25,7 +25,8 @@ run_hook = lang_base.basic_run_hook
 
 
 def _resource_bytesio(filename: str) -> IO[bytes]:
-    return importlib.resources.open_binary('pre_commit.resources', filename)
+    files = importlib.resources.files('pre_commit.resources')
+    return files.joinpath(filename).open('rb')
 
 
 @functools.lru_cache(maxsize=1)
@@ -72,7 +73,7 @@ def get_env_patch(
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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
@@ -114,6 +115,8 @@ def _install_ruby(
 def install_environment(
         prefix: Prefix, version: str, additional_dependencies: Sequence[str],
 ) -> None:
+    envdir = lang_base.environment_dir(prefix, ENVIRONMENT_DIR, version)
+
     if version != 'system':  # pragma: win32 no cover
         _install_rbenv(prefix, version)
         with in_env(prefix, version):
@@ -135,6 +138,8 @@ def install_environment(
                 '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,
             ),
         )

pre_commit/languages/rust.py

@@ -7,8 +7,8 @@ import shutil
 import sys
 import tempfile
 import urllib.request
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 import pre_commit.constants as C
 from pre_commit import lang_base
@@ -34,7 +34,7 @@ def get_default_version() -> str:
     # 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)[0] == 0:
+    if cmd_output_b('cargo', '--version', check=False, cwd='/')[0] == 0:
         return 'system'
     else:
         return C.DEFAULT
@@ -61,7 +61,7 @@ def get_env_patch(target_dir: str, version: str) -> PatchesT:
 
 
 @contextlib.contextmanager
-def in_env(prefix: Prefix, version: str) -> Generator[None, None, None]:
+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
@@ -134,7 +134,7 @@ def install_environment(
 
     packages_to_install: set[tuple[str, ...]] = {('--path', '.')}
     for cli_dep in cli_deps:
-        cli_dep = cli_dep[len('cli:'):]
+        cli_dep = cli_dep.removeprefix('cli:')
         package, _, crate_version = cli_dep.partition(':')
         if crate_version != '':
             packages_to_install.add((package, '--version', crate_version))

pre_commit/languages/swift.py

@@ -2,8 +2,8 @@ from __future__ import annotations
 
 import contextlib
 import os
-from typing import Generator
-from typing import Sequence
+from collections.abc import Generator
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.envcontext import envcontext
@@ -27,7 +27,7 @@ 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, None, None]:
+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

pre_commit/languages/unsupported_script.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit import lang_base
 from pre_commit.prefix import Prefix

pre_commit/logging_handler.py

@@ -2,7 +2,7 @@ from __future__ import annotations
 
 import contextlib
 import logging
-from typing import Generator
+from collections.abc import Generator
 
 from pre_commit import color
 from pre_commit import output
@@ -32,7 +32,7 @@ class LoggingHandler(logging.Handler):
 
 
 @contextlib.contextmanager
-def logging_handler(use_color: bool) -> Generator[None, None, None]:
+def logging_handler(use_color: bool) -> Generator[None]:
     handler = LoggingHandler(use_color)
     logger.addHandler(handler)
     logger.setLevel(logging.INFO)

pre_commit/main.py

@@ -4,12 +4,13 @@ import argparse
 import logging
 import os
 import sys
-from typing import Sequence
+from collections.abc import Sequence
 
 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
@@ -37,8 +38,11 @@ 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', 'init-templatedir', 'sample-config',
+    'clean', 'gc', 'hazmat', 'init-templatedir', 'sample-config',
     'validate-config', 'validate-manifest',
 }
 
@@ -59,10 +63,10 @@ def _add_hook_type_option(parser: argparse.ArgumentParser) -> None:
 
 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', default=False)
+    parser.add_argument('--verbose', '-v', action='store_true')
     mutex_group = parser.add_mutually_exclusive_group(required=False)
     mutex_group.add_argument(
-        '--all-files', '-a', action='store_true', default=False,
+        '--all-files', '-a', action='store_true',
         help='Run on all the files in the repo.',
     )
     mutex_group.add_argument(
@@ -73,6 +77,10 @@ def _add_run_options(parser: argparse.ArgumentParser) -> None:
         '--show-diff-on-failure', action='store_true',
         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,
@@ -238,6 +246,11 @@ def main(argv: Sequence[str] | None = None) -> int:
 
     _add_cmd('gc', help='Clean unused cached repos.')
 
+    hazmat_parser = _add_cmd(
+        'hazmat', help='Composable tools for rare use in hook `entry`.',
+    )
+    hazmat.add_parsers(hazmat_parser)
+
     init_templatedir_parser = _add_cmd(
         'init-templatedir',
         help=(
@@ -272,7 +285,7 @@ def main(argv: Sequence[str] | None = None) -> int:
     )
     _add_hook_type_option(install_parser)
     install_parser.add_argument(
-        '--allow-missing-config', action='store_true', default=False,
+        '--allow-missing-config', action='store_true',
         help=(
             'Whether to allow a missing `pre-commit` configuration file '
             'or exit with a failure code.'
@@ -382,6 +395,8 @@ def main(argv: Sequence[str] | None = None) -> int:
             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,

pre_commit/meta_hooks/check_hooks_apply.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import argparse
-from typing import Sequence
+from collections.abc import Sequence
 
 import pre_commit.constants as C
 from pre_commit import git
@@ -21,7 +21,7 @@ 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 classifier.filenames_for_hook(hook):
+        elif not any(classifier.filenames_for_hook(hook)):
             print(f'{hook.id} does not apply to this repository')
             retv = 1
 

pre_commit/meta_hooks/check_useless_excludes.py

@@ -2,7 +2,8 @@ from __future__ import annotations
 
 import argparse
 import re
-from typing import Sequence
+from collections.abc import Iterable
+from collections.abc import Sequence
 
 from cfgv import apply_defaults
 
@@ -14,7 +15,7 @@ from pre_commit.commands.run import Classifier
 
 
 def exclude_matches_any(
-        filenames: Sequence[str],
+        filenames: Iterable[str],
         include: str,
         exclude: str,
 ) -> bool:
@@ -50,11 +51,12 @@ 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.filenames
-            types = hook['types']
-            types_or = hook['types_or']
-            exclude_types = hook['exclude_types']
-            names = classifier.by_types(names, types, types_or, exclude_types)
+            names = classifier.by_types(
+                classifier.filenames,
+                hook['types'],
+                hook['types_or'],
+                hook['exclude_types'],
+            )
             include, exclude = hook['files'], hook['exclude']
             if not exclude_matches_any(names, include, exclude):
                 print(

pre_commit/meta_hooks/identity.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import sys
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit import output
 

pre_commit/parse_shebang.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import os.path
-from typing import Mapping
+from collections.abc import Mapping
 from typing import NoReturn
 
 from identify.identify import parse_shebang_from_file

pre_commit/repository.py

@@ -3,16 +3,14 @@ from __future__ import annotations
 import json
 import logging
 import os
-import shlex
+from collections.abc import Sequence
 from typing import Any
-from typing import Sequence
 
 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.clientlib import parse_version
 from pre_commit.hook import Hook
 from pre_commit.lang_base import environment_dir
 from pre_commit.prefix import Prefix
@@ -69,14 +67,6 @@ def _hook_install(hook: Hook) -> None:
     logger.info('Once installed this environment will be reused.')
     logger.info('This may take a few minutes...')
 
-    if hook.language == 'python_venv':
-        logger.warning(
-            f'`repo: {hook.src}` uses deprecated `language: python_venv`.  '
-            f'This is an alias for `language: python`.  '
-            f'Often `pre-commit autoupdate --repo {shlex.quote(hook.src)}` '
-            f'will fix this.',
-        )
-
     lang = languages[hook.language]
     assert lang.ENVIRONMENT_DIR is not None
 
@@ -124,15 +114,6 @@ def _hook(
     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]

pre_commit/resources/empty_template_pubspec.yaml

@@ -1,4 +1,4 @@
 name: pre_commit_empty_pubspec
 environment:
-  sdk: '>=2.10.0'
+  sdk: '>=2.12.0'
 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')
+setup(name='pre-commit-placeholder-package', version='0.0.0', py_modules=[])

pre_commit/staged_files_only.py

@@ -4,7 +4,7 @@ import contextlib
 import logging
 import os.path
 import time
-from typing import Generator
+from collections.abc import Generator
 
 from pre_commit import git
 from pre_commit.errors import FatalError
@@ -33,7 +33,7 @@ def _git_apply(patch: str) -> None:
 
 
 @contextlib.contextmanager
-def _intent_to_add_cleared() -> Generator[None, None, None]:
+def _intent_to_add_cleared() -> Generator[None]:
     intent_to_add = git.intent_to_add_files()
     if intent_to_add:
         logger.warning('Unstaged intent-to-add files detected.')
@@ -48,7 +48,7 @@ def _intent_to_add_cleared() -> Generator[None, None, None]:
 
 
 @contextlib.contextmanager
-def _unstaged_changes_cleared(patch_dir: str) -> Generator[None, None, None]:
+def _unstaged_changes_cleared(patch_dir: str) -> Generator[None]:
     tree = cmd_output('git', 'write-tree')[1].strip()
     diff_cmd = (
         'git', 'diff-index', '--ignore-submodules', '--binary',
@@ -59,6 +59,11 @@ def _unstaged_changes_cleared(patch_dir: str) -> Generator[None, None, None]:
         # 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():
         patch_filename = f'patch{int(time.time())}-{os.getpid()}'
         patch_filename = os.path.join(patch_dir, patch_filename)
@@ -100,7 +105,7 @@ def _unstaged_changes_cleared(patch_dir: str) -> Generator[None, None, None]:
 
 
 @contextlib.contextmanager
-def staged_files_only(patch_dir: str) -> Generator[None, None, None]:
+def staged_files_only(patch_dir: str) -> Generator[None]:
     """Clear any unstaged changes from the git working directory inside this
     context.
     """

pre_commit/store.py

@@ -5,18 +5,18 @@ import logging
 import os.path
 import sqlite3
 import tempfile
-from typing import Callable
-from typing import Generator
-from typing import Sequence
+from collections.abc import Callable
+from collections.abc import Generator
+from collections.abc import Sequence
 
 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')
@@ -95,13 +95,13 @@ class Store:
                     '    PRIMARY KEY (repo, ref)'
                     ');',
                 )
-                self._create_config_table(db)
+                self._create_configs_table(db)
 
             # Atomic file move
             os.replace(tmpfile, self.db_path)
 
     @contextlib.contextmanager
-    def exclusive_lock(self) -> Generator[None, None, None]:
+    def exclusive_lock(self) -> Generator[None]:
         def blocked_cb() -> None:  # pragma: no cover (tests are in-process)
             logger.info('Locking pre-commit directory')
 
@@ -112,7 +112,7 @@ class Store:
     def connect(
             self,
             db_path: str | None = None,
-    ) -> Generator[sqlite3.Connection, None, None]:
+    ) -> Generator[sqlite3.Connection]:
         db_path = db_path or self.db_path
         # sqlite doesn't close its fd with its contextmanager >.<
         # contextlib.closing fixes this.
@@ -136,6 +136,7 @@ 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:
@@ -168,6 +169,9 @@ 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:
@@ -210,7 +214,7 @@ class Store:
             'local', C.LOCAL_REPO_VERSION, deps, _make_local_repo,
         )
 
-    def _create_config_table(self, db: sqlite3.Connection) -> None:
+    def _create_configs_table(self, db: sqlite3.Connection) -> None:
         db.executescript(
             'CREATE TABLE IF NOT EXISTS configs ('
             '   path TEXT NOT NULL,'
@@ -227,28 +231,5 @@ class Store:
             return
         with self.connect() as db:
             # TODO: eventually remove this and only create in _create
-            self._create_config_table(db)
+            self._create_configs_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

@@ -8,10 +8,10 @@ import shutil
 import stat
 import subprocess
 import sys
+from collections.abc import Callable
+from collections.abc import Generator
 from types import TracebackType
 from typing import Any
-from typing import Callable
-from typing import Generator
 
 from pre_commit import parse_shebang
 
@@ -25,7 +25,7 @@ def force_bytes(exc: Any) -> bytes:
 
 
 @contextlib.contextmanager
-def clean_path_on_failure(path: str) -> Generator[None, None, None]:
+def clean_path_on_failure(path: str) -> Generator[None]:
     """Cleans up the directory on an exceptional failure."""
     try:
         yield
@@ -36,7 +36,8 @@ def clean_path_on_failure(path: str) -> Generator[None, None, None]:
 
 
 def resource_text(filename: str) -> str:
-    return importlib.resources.read_text('pre_commit.resources', filename)
+    files = importlib.resources.files('pre_commit.resources')
+    return files.joinpath(filename).read_text()
 
 
 def make_executable(filename: str) -> None:
@@ -201,24 +202,37 @@ else:  # pragma: no cover
     cmd_output_p = cmd_output_b
 
 
-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],
+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],
     ) -> None:
-        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)
+        return _handle_readonly(func, path, excinfo[1])
+
+    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 win_exe(s: str) -> str:

pre_commit/xargs.py

@@ -7,12 +7,12 @@ 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 MutableMapping
-from typing import Sequence
 from typing import TypeVar
 
 from pre_commit import parse_shebang
@@ -24,6 +24,14 @@ 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:
@@ -112,7 +120,6 @@ 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
@@ -170,7 +177,8 @@ def xargs(
         results = thread_map(run_cmd_partition, partitions)
 
         for proc_retcode, proc_out, _ in results:
-            retcode = max(retcode, proc_retcode)
+            if abs(proc_retcode) > abs(retcode):
+                retcode = proc_retcode
             stdout += proc_out
 
     return retcode, stdout

pre_commit/yaml.py

@@ -6,6 +6,7 @@ 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)
 

pre_commit/yaml_rewrite.py

@@ -0,0 +1,52 @@
+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))

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = pre_commit
-version = 3.3.0
+version = 4.5.1
 description = A framework for managing and maintaining multi-language pre-commit hooks.
 long_description = file: README.md
 long_description_content_type = text/markdown
@@ -8,9 +8,8 @@ url = https://github.com/pre-commit/pre-commit
 author = Anthony Sottile
 author_email = asottile@umich.edu
 license = MIT
-license_file = LICENSE
+license_files = LICENSE
 classifiers =
-    License :: OSI Approved :: MIT License
     Programming Language :: Python :: 3
     Programming Language :: Python :: 3 :: Only
     Programming Language :: Python :: Implementation :: CPython
@@ -24,7 +23,7 @@ install_requires =
     nodeenv>=0.11.1
     pyyaml>=5.1
     virtualenv>=20.10.0
-python_requires = >=3.8
+python_requires = >=3.10
 
 [options.packages.find]
 exclude =
@@ -53,6 +52,7 @@ check_untyped_defs = true
 disallow_any_generics = true
 disallow_incomplete_defs = true
 disallow_untyped_defs = true
+enable_error_code = deprecated
 warn_redundant_casts = true
 warn_unused_ignores = true
 

testing/get-dart.sh

@@ -1,7 +1,7 @@
 #!/usr/bin/env bash
 set -euo pipefail
 
-VERSION=2.13.4
+VERSION=2.19.6
 
 if [ "$OSTYPE" = msys ]; then
     URL="https://storage.googleapis.com/dart-archive/channels/stable/release/${VERSION}/sdk/dartsdk-windows-x64-release.zip"

testing/get-swift.sh

@@ -1,29 +0,0 @@
-#!/usr/bin/env bash
-# This is a script used in CI to install swift
-set -euo pipefail
-
-. /etc/lsb-release
-if [ "$DISTRIB_CODENAME" = "jammy" ]; then
-    SWIFT_URL='https://download.swift.org/swift-5.7.1-release/ubuntu2204/swift-5.7.1-RELEASE/swift-5.7.1-RELEASE-ubuntu22.04.tar.gz'
-    SWIFT_HASH='7f60291f5088d3e77b0c2364beaabd29616ee7b37260b7b06bdbeb891a7fe161'
-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 '/tmp/swift/usr/bin' >> "$GITHUB_PATH"

testing/language_helpers.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import os
-from typing import Sequence
+from collections.abc import Sequence
 
 from pre_commit.lang_base import Language
 from pre_commit.prefix import Prefix

testing/languages

@@ -16,6 +16,15 @@ EXCLUDED = frozenset((
 ))
 
 
+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
@@ -47,10 +56,14 @@ def main() -> int:
         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
+                lang: files | triggers_all
                 for lang, files in zip(langs, exe.map(_lang_files, langs))
             }
 

testing/make-archives

@@ -8,7 +8,7 @@ import shutil
 import subprocess
 import tarfile
 import tempfile
-from typing import Sequence
+from collections.abc import Sequence
 
 
 # This is a script for generating the tarred resources for git repo
@@ -16,8 +16,8 @@ from typing import Sequence
 
 
 REPOS = (
-    ('rbenv', 'https://github.com/rbenv/rbenv', '38e1fbb'),
-    ('ruby-build', 'https://github.com/rbenv/ruby-build', '855b963'),
+    ('rbenv', 'https://github.com/rbenv/rbenv', '10e96bfc'),
+    ('ruby-build', 'https://github.com/rbenv/ruby-build', '447468b1'),
     (
         'ruby-download',
         'https://github.com/garnieretienne/rvm-download',
@@ -57,8 +57,7 @@ def make_archive(name: str, repo: str, ref: str, destdir: str) -> str:
         arcs.sort()
 
         with gzip.GzipFile(output_path, 'wb', mtime=0) as gzipf:
-            # https://github.com/python/typeshed/issues/5491
-            with tarfile.open(fileobj=gzipf, mode='w') as tf:  # type: ignore
+            with tarfile.open(fileobj=gzipf, mode='w') as tf:
                 for arcname, abspath in arcs:
                     tf.add(
                         abspath,

testing/resources/python3_hooks_repo/.pre-commit-hooks.yaml

@@ -1,6 +0,0 @@
--   id: python3-hook
-    name: Python 3 Hook
-    entry: python3-hook
-    language: python
-    language_version: python3
-    files: \.py$

testing/resources/python3_hooks_repo/py3_hook.py

@@ -1,8 +0,0 @@
-import sys
-
-
-def main():
-    print(sys.version_info[0])
-    print(repr(sys.argv[1:]))
-    print('Hello World')
-    return 0

testing/resources/python3_hooks_repo/setup.py

@@ -1,8 +0,0 @@
-from setuptools import setup
-
-setup(
-    name='python3_hook',
-    version='0.0.0',
-    py_modules=['py3_hook'],
-    entry_points={'console_scripts': ['python3-hook = py3_hook:main']},
-)

testing/resources/system_hook_with_spaces_repo/.pre-commit-hooks.yaml

@@ -1,5 +0,0 @@
--   id: system-hook-with-spaces
-    name: System hook with spaces
-    entry: bash -c 'echo "Hello World"'
-    language: system
-    files: \.sh$

testing/util.py

@@ -40,6 +40,7 @@ def run_opts(
         color=False,
         verbose=False,
         hook=None,
+        fail_fast=False,
         remote_branch='',
         local_branch='',
         from_ref='',
@@ -65,6 +66,7 @@ def run_opts(
         color=color,
         verbose=verbose,
         hook=hook,
+        fail_fast=fail_fast,
         remote_branch=remote_branch,
         local_branch=local_branch,
         from_ref=from_ref,

testing/zipapp/Dockerfile

@@ -1,4 +1,4 @@
-FROM ubuntu:focal
+FROM ubuntu:jammy
 RUN : \
     && apt-get update \
     && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
@@ -11,4 +11,4 @@ RUN : \
 ENV LANG=C.UTF-8 PATH=/venv/bin:$PATH
 RUN : \
     && python3 -mvenv /venv \
-    && pip install --no-cache-dir pip setuptools wheel no-manylinux --upgrade
+    && pip install --no-cache-dir pip distlib no-manylinux --upgrade

testing/zipapp/make

@@ -4,7 +4,6 @@ from __future__ import annotations
 import argparse
 import base64
 import hashlib
-import importlib.resources
 import io
 import os.path
 import shutil
@@ -42,10 +41,17 @@ def _add_shim(dest: str) -> None:
     with zipfile.ZipFile(bio, 'w') as zipf:
         zipf.write(shim, arcname='__main__.py')
 
-    with open(os.path.join(dest, 'python.exe'), 'wb') as f:
-        f.write(importlib.resources.read_binary('distlib', 't32.exe'))
-        f.write(b'#!py.exe -3\n')
-        f.write(bio.getvalue())
+    with tempfile.TemporaryDirectory() as tmpdir:
+        _exit_if_retv(
+            'podman', 'run', '--rm', '--volume', f'{tmpdir}:/out:rw', IMG,
+            'cp', '/venv/lib/python3.10/site-packages/distlib/t32.exe', '/out',
+        )
+
+        with open(os.path.join(dest, 'python.exe'), 'wb') as f:
+            with open(os.path.join(tmpdir, 't32.exe'), 'rb') as t32:
+                f.write(t32.read())
+            f.write(b'#!py.exe -3\n')
+            f.write(bio.getvalue())
 
 
 def _write_cache_key(version: str, wheeldir: str, dest: str) -> None:
@@ -101,9 +107,6 @@ def main() -> int:
         shebang = '/usr/bin/env python3'
         zipapp.create_archive(tmpdir, filename, interpreter=shebang)
 
-    with open(f'{filename}.sha256sum', 'w') as f:
-        subprocess.check_call(('sha256sum', filename), stdout=f)
-
     return 0
 
 

tests/all_languages_test.py

@@ -1,7 +0,0 @@
-from __future__ import annotations
-
-from pre_commit.all_languages import languages
-
-
-def test_python_venv_is_an_alias_to_python():
-    assert languages['python_venv'] is languages['python']

tests/clientlib_test.py

@@ -12,6 +12,8 @@ from pre_commit.clientlib import CONFIG_HOOK_DICT
 from pre_commit.clientlib import CONFIG_REPO_DICT
 from pre_commit.clientlib import CONFIG_SCHEMA
 from pre_commit.clientlib import DEFAULT_LANGUAGE_VERSION
+from pre_commit.clientlib import InvalidManifestError
+from pre_commit.clientlib import load_manifest
 from pre_commit.clientlib import MANIFEST_HOOK_DICT
 from pre_commit.clientlib import MANIFEST_SCHEMA
 from pre_commit.clientlib import META_HOOK_DICT
@@ -40,56 +42,51 @@ def test_check_type_tag_success():
 
 
 @pytest.mark.parametrize(
-    ('config_obj', 'expected'), (
-        (
-            {
-                'repos': [{
-                    'repo': 'git@github.com:pre-commit/pre-commit-hooks',
-                    'rev': 'cd74dc150c142c3be70b24eaf0b02cae9d235f37',
-                    'hooks': [{'id': 'pyflakes', 'files': '\\.py$'}],
-                }],
-            },
-            True,
-        ),
-        (
-            {
-                'repos': [{
-                    'repo': 'git@github.com:pre-commit/pre-commit-hooks',
-                    'rev': 'cd74dc150c142c3be70b24eaf0b02cae9d235f37',
-                    'hooks': [
-                        {
-                            'id': 'pyflakes',
-                            'files': '\\.py$',
-                            'args': ['foo', 'bar', 'baz'],
-                        },
-                    ],
-                }],
-            },
-            True,
-        ),
-        (
-            {
-                'repos': [{
-                    'repo': 'git@github.com:pre-commit/pre-commit-hooks',
-                    'rev': 'cd74dc150c142c3be70b24eaf0b02cae9d235f37',
-                    'hooks': [
-                        {
-                            'id': 'pyflakes',
-                            'files': '\\.py$',
-                            # Exclude pattern must be a string
-                            'exclude': 0,
-                            'args': ['foo', 'bar', 'baz'],
-                        },
-                    ],
-                }],
-            },
-            False,
-        ),
+    'cfg',
+    (
+        {
+            'repos': [{
+                'repo': 'git@github.com:pre-commit/pre-commit-hooks',
+                'rev': 'cd74dc150c142c3be70b24eaf0b02cae9d235f37',
+                'hooks': [{'id': 'pyflakes', 'files': '\\.py$'}],
+            }],
+        },
+        {
+            'repos': [{
+                'repo': 'git@github.com:pre-commit/pre-commit-hooks',
+                'rev': 'cd74dc150c142c3be70b24eaf0b02cae9d235f37',
+                'hooks': [
+                    {
+                        'id': 'pyflakes',
+                        'files': '\\.py$',
+                        'args': ['foo', 'bar', 'baz'],
+                    },
+                ],
+            }],
+        },
     ),
 )
-def test_config_valid(config_obj, expected):
-    ret = is_valid_according_to_schema(config_obj, CONFIG_SCHEMA)
-    assert ret is expected
+def test_config_valid(cfg):
+    assert is_valid_according_to_schema(cfg, CONFIG_SCHEMA)
+
+
+def test_invalid_config_wrong_type():
+    cfg = {
+        'repos': [{
+            'repo': 'git@github.com:pre-commit/pre-commit-hooks',
+            'rev': 'cd74dc150c142c3be70b24eaf0b02cae9d235f37',
+            'hooks': [
+                {
+                    'id': 'pyflakes',
+                    'files': '\\.py$',
+                    # Exclude pattern must be a string
+                    'exclude': 0,
+                    'args': ['foo', 'bar', 'baz'],
+                },
+            ],
+        }],
+    }
+    assert not is_valid_according_to_schema(cfg, CONFIG_SCHEMA)
 
 
 def test_local_hooks_with_rev_fails():
@@ -198,14 +195,13 @@ def test_warn_mutable_rev_conditional():
     ),
 )
 def test_sensible_regex_validators_dont_pass_none(validator_cls):
-    key = 'files'
+    validator = validator_cls('files', cfgv.check_string)
     with pytest.raises(cfgv.ValidationError) as excinfo:
-        validator = validator_cls(key, cfgv.check_string)
-        validator.check({key: None})
+        validator.check({'files': None})
 
     assert str(excinfo.value) == (
         '\n'
-        f'==> At key: {key}'
+        '==> At key: files'
         '\n'
         '=====> Expected string got NoneType'
     )
@@ -262,6 +258,24 @@ def test_validate_optional_sensible_regex_at_local_hook(caplog):
     ]
 
 
+def test_validate_optional_sensible_regex_at_meta_hook(caplog):
+    config_obj = {
+        'repo': 'meta',
+        'hooks': [{'id': 'identity', 'files': 'dir/*.py'}],
+    }
+
+    cfgv.validate(config_obj, CONFIG_REPO_DICT)
+
+    assert caplog.record_tuples == [
+        (
+            'pre_commit',
+            logging.WARNING,
+            "The 'files' field in hook 'identity' is a regex, not a glob "
+            "-- matching '/*' probably isn't what you want here",
+        ),
+    ]
+
+
 @pytest.mark.parametrize(
     ('regex', 'warning'),
     (
@@ -297,47 +311,128 @@ def test_validate_optional_sensible_regex_at_top_level(caplog, regex, warning):
     assert caplog.record_tuples == [('pre_commit', logging.WARNING, warning)]
 
 
+def test_invalid_stages_error():
+    cfg = {'repos': [sample_local_config()]}
+    cfg['repos'][0]['hooks'][0]['stages'] = ['invalid']
+
+    with pytest.raises(cfgv.ValidationError) as excinfo:
+        cfgv.validate(cfg, CONFIG_SCHEMA)
+
+    assert str(excinfo.value) == (
+        '\n'
+        '==> At Config()\n'
+        '==> At key: repos\n'
+        "==> At Repository(repo='local')\n"
+        '==> At key: hooks\n'
+        "==> At Hook(id='do_not_commit')\n"
+        # this line was missing due to the custom validator
+        '==> At key: stages\n'
+        '==> At index 0\n'
+        "=====> Expected one of commit-msg, manual, post-checkout, post-commit, post-merge, post-rewrite, pre-commit, pre-merge-commit, pre-push, pre-rebase, prepare-commit-msg but got: 'invalid'"  # noqa: E501
+    )
+
+
+def test_warning_for_deprecated_stages(caplog):
+    config_obj = sample_local_config()
+    config_obj['hooks'][0]['stages'] = ['commit', 'push']
+
+    cfgv.validate(config_obj, CONFIG_REPO_DICT)
+
+    assert caplog.record_tuples == [
+        (
+            'pre_commit',
+            logging.WARNING,
+            'hook id `do_not_commit` uses deprecated stage names '
+            '(commit, push) which will be removed in a future version.  '
+            'run: `pre-commit migrate-config` to automatically fix this.',
+        ),
+    ]
+
+
+def test_no_warning_for_non_deprecated_stages(caplog):
+    config_obj = sample_local_config()
+    config_obj['hooks'][0]['stages'] = ['pre-commit', 'pre-push']
+
+    cfgv.validate(config_obj, CONFIG_REPO_DICT)
+
+    assert caplog.record_tuples == []
+
+
+def test_warning_for_deprecated_default_stages(caplog):
+    cfg = {'default_stages': ['commit', 'push'], 'repos': []}
+
+    cfgv.validate(cfg, CONFIG_SCHEMA)
+
+    assert caplog.record_tuples == [
+        (
+            'pre_commit',
+            logging.WARNING,
+            'top-level `default_stages` uses deprecated stage names '
+            '(commit, push) which will be removed in a future version.  '
+            'run: `pre-commit migrate-config` to automatically fix this.',
+        ),
+    ]
+
+
+def test_no_warning_for_non_deprecated_default_stages(caplog):
+    cfg = {'default_stages': ['pre-commit', 'pre-push'], 'repos': []}
+
+    cfgv.validate(cfg, CONFIG_SCHEMA)
+
+    assert caplog.record_tuples == []
+
+
+def test_unsupported_language_migration():
+    cfg = {'repos': [sample_local_config(), sample_local_config()]}
+    cfg['repos'][0]['hooks'][0]['language'] = 'system'
+    cfg['repos'][1]['hooks'][0]['language'] = 'script'
+
+    cfgv.validate(cfg, CONFIG_SCHEMA)
+    ret = cfgv.apply_defaults(cfg, CONFIG_SCHEMA)
+
+    assert ret['repos'][0]['hooks'][0]['language'] == 'unsupported'
+    assert ret['repos'][1]['hooks'][0]['language'] == 'unsupported_script'
+
+
+def test_unsupported_language_migration_language_required():
+    cfg = {'repos': [sample_local_config()]}
+    del cfg['repos'][0]['hooks'][0]['language']
+
+    with pytest.raises(cfgv.ValidationError):
+        cfgv.validate(cfg, CONFIG_SCHEMA)
+
+
 @pytest.mark.parametrize(
-    ('manifest_obj', 'expected'),
+    'manifest_obj',
     (
-        (
-            [{
-                'id': 'a',
-                'name': 'b',
-                'entry': 'c',
-                'language': 'python',
-                'files': r'\.py$',
-            }],
-            True,
-        ),
-        (
-            [{
-                'id': 'a',
-                'name': 'b',
-                'entry': 'c',
-                'language': 'python',
-                'language_version': 'python3.4',
-                'files': r'\.py$',
-            }],
-            True,
-        ),
-        (
-            # A regression in 0.13.5: always_run and files are permissible
-            [{
-                'id': 'a',
-                'name': 'b',
-                'entry': 'c',
-                'language': 'python',
-                'files': '',
-                'always_run': True,
-            }],
-            True,
-        ),
+        [{
+            'id': 'a',
+            'name': 'b',
+            'entry': 'c',
+            'language': 'python',
+            'files': r'\.py$',
+        }],
+        [{
+            'id': 'a',
+            'name': 'b',
+            'entry': 'c',
+            'language': 'python',
+            'language_version': 'python3.4',
+            'files': r'\.py$',
+        }],
+        # A regression in 0.13.5: always_run and files are permissible
+        [{
+            'id': 'a',
+            'name': 'b',
+            'entry': 'c',
+            'language': 'python',
+            'files': '',
+            'always_run': True,
+        }],
     ),
 )
-def test_valid_manifests(manifest_obj, expected):
-    ret = is_valid_according_to_schema(manifest_obj, MANIFEST_SCHEMA)
-    assert ret is expected
+def test_valid_manifests(manifest_obj):
+    assert is_valid_according_to_schema(manifest_obj, MANIFEST_SCHEMA)
 
 
 @pytest.mark.parametrize(
@@ -393,8 +488,39 @@ def test_parse_version():
 
 
 def test_minimum_pre_commit_version_failing():
+    cfg = {'repos': [], 'minimum_pre_commit_version': '999'}
+    with pytest.raises(cfgv.ValidationError) as excinfo:
+        cfgv.validate(cfg, CONFIG_SCHEMA)
+    assert str(excinfo.value) == (
+        f'\n'
+        f'==> At Config()\n'
+        f'==> At key: minimum_pre_commit_version\n'
+        f'=====> pre-commit version 999 is required but version {C.VERSION} '
+        f'is installed.  Perhaps run `pip install --upgrade pre-commit`.'
+    )
+
+
+def test_minimum_pre_commit_version_failing_in_config():
+    cfg = {'repos': [sample_local_config()]}
+    cfg['repos'][0]['hooks'][0]['minimum_pre_commit_version'] = '999'
+    with pytest.raises(cfgv.ValidationError) as excinfo:
+        cfgv.validate(cfg, CONFIG_SCHEMA)
+    assert str(excinfo.value) == (
+        f'\n'
+        f'==> At Config()\n'
+        f'==> At key: repos\n'
+        f"==> At Repository(repo='local')\n"
+        f'==> At key: hooks\n'
+        f"==> At Hook(id='do_not_commit')\n"
+        f'==> At key: minimum_pre_commit_version\n'
+        f'=====> pre-commit version 999 is required but version {C.VERSION} '
+        f'is installed.  Perhaps run `pip install --upgrade pre-commit`.'
+    )
+
+
+def test_minimum_pre_commit_version_failing_before_other_error():
+    cfg = {'repos': 5, 'minimum_pre_commit_version': '999'}
     with pytest.raises(cfgv.ValidationError) as excinfo:
-        cfg = {'repos': [], 'minimum_pre_commit_version': '999'}
         cfgv.validate(cfg, CONFIG_SCHEMA)
     assert str(excinfo.value) == (
         f'\n'
@@ -464,3 +590,18 @@ def test_config_hook_stages_defaulting():
         'id': 'fake-hook',
         'stages': ['commit-msg', 'pre-push', 'pre-commit', 'pre-merge-commit'],
     }
+
+
+def test_manifest_v5_forward_compat(tmp_path):
+    manifest = tmp_path.joinpath('.pre-commit-hooks.yaml')
+    manifest.write_text('hooks: {}')
+
+    with pytest.raises(InvalidManifestError) as excinfo:
+        load_manifest(manifest)
+    assert str(excinfo.value) == (
+        f'\n'
+        f'==> File {manifest}\n'
+        f'=====> \n'
+        f'=====> pre-commit version 5 is required but version {C.VERSION} '
+        f'is installed.  Perhaps run `pip install --upgrade pre-commit`.'
+    )

tests/commands/gc_test.py

@@ -19,11 +19,13 @@ from testing.util import git_commit
 
 
 def _repo_count(store):
-    return len(store.select_all_repos())
+    with store.connect() as db:
+        return db.execute('SELECT COUNT(1) FROM repos').fetchone()[0]
 
 
 def _config_count(store):
-    return len(store.select_all_configs())
+    with store.connect() as db:
+        return db.execute('SELECT COUNT(1) FROM configs').fetchone()[0]
 
 
 def _remove_config_assert_cleared(store, cap_out):
@@ -153,7 +155,8 @@ def test_invalid_manifest_gcd(tempdir_factory, store, in_git_dir, cap_out):
     install_hooks(C.CONFIG_FILE, store)
 
     # we'll "break" the manifest to simulate an old version clone
-    (_, _, path), = store.select_all_repos()
+    with store.connect() as db:
+        path, = db.execute('SELECT path FROM repos').fetchone()
     os.remove(os.path.join(path, C.MANIFEST_FILE))
 
     assert _config_count(store) == 1
@@ -162,3 +165,11 @@ def test_invalid_manifest_gcd(tempdir_factory, store, in_git_dir, cap_out):
     assert _config_count(store) == 1
     assert _repo_count(store) == 0
     assert cap_out.get().splitlines()[-1] == '1 repo(s) removed.'
+
+
+def test_gc_pre_1_14_roll_forward(store, cap_out):
+    with store.connect() as db:  # simulate pre-1.14.0
+        db.executescript('DROP TABLE configs')
+
+    assert not gc(store)
+    assert cap_out.get() == '0 repo(s) removed.\n'

tests/commands/hazmat_test.py

@@ -0,0 +1,99 @@
+from __future__ import annotations
+
+import sys
+
+import pytest
+
+from pre_commit.commands.hazmat import _cmd_filenames
+from pre_commit.commands.hazmat import main
+from testing.util import cwd
+
+
+def test_cmd_filenames_no_dash_dash():
+    with pytest.raises(SystemExit) as excinfo:
+        _cmd_filenames(('no', 'dashdash', 'here'))
+    msg, = excinfo.value.args
+    assert msg == 'hazmat entry must end with `--`'
+
+
+def test_cmd_filenames_no_filenames():
+    cmd, filenames = _cmd_filenames(('hello', 'world', '--'))
+    assert cmd == ('hello', 'world')
+    assert filenames == ()
+
+
+def test_cmd_filenames_some_filenames():
+    cmd, filenames = _cmd_filenames(('hello', 'world', '--', 'f1', 'f2'))
+    assert cmd == ('hello', 'world')
+    assert filenames == ('f1', 'f2')
+
+
+def test_cmd_filenames_multiple_dashdash():
+    cmd, filenames = _cmd_filenames(('hello', '--', 'arg', '--', 'f1', 'f2'))
+    assert cmd == ('hello', '--', 'arg')
+    assert filenames == ('f1', 'f2')
+
+
+def test_cd_unexpected_filename():
+    with pytest.raises(SystemExit) as excinfo:
+        main(('cd', 'subdir', 'cmd', '--', 'subdir/1', 'not-subdir/2'))
+    msg, = excinfo.value.args
+    assert msg == "unexpected file without prefix='subdir/': not-subdir/2"
+
+
+def _norm(out):
+    return out.replace('\r\n', '\n')
+
+
+def test_cd(tmp_path, capfd):
+    subdir = tmp_path.joinpath('subdir')
+    subdir.mkdir()
+    subdir.joinpath('a').write_text('a')
+    subdir.joinpath('b').write_text('b')
+
+    with cwd(tmp_path):
+        ret = main((
+            'cd', 'subdir',
+            sys.executable, '-c',
+            'import os; print(os.getcwd());'
+            'import sys; [print(open(f).read()) for f in sys.argv[1:]]',
+            '--',
+            'subdir/a', 'subdir/b',
+        ))
+
+    assert ret == 0
+    out, err = capfd.readouterr()
+    assert _norm(out) == f'{subdir}\na\nb\n'
+    assert err == ''
+
+
+def test_ignore_exit_code(capfd):
+    ret = main((
+        'ignore-exit-code', sys.executable, '-c', 'raise SystemExit("bye")',
+    ))
+    assert ret == 0
+    out, err = capfd.readouterr()
+    assert out == ''
+    assert _norm(err) == 'bye\n'
+
+
+def test_n1(capfd):
+    ret = main((
+        'n1', sys.executable, '-c', 'import sys; print(sys.argv[1:])',
+        '--',
+        'foo', 'bar', 'baz',
+    ))
+    assert ret == 0
+    out, err = capfd.readouterr()
+    assert _norm(out) == "['foo']\n['bar']\n['baz']\n"
+    assert err == ''
+
+
+def test_n1_some_error_code():
+    ret = main((
+        'n1', sys.executable, '-c',
+        'import sys; raise SystemExit(sys.argv[1] == "error")',
+        '--',
+        'ok', 'error', 'ok',
+    ))
+    assert ret == 1

tests/commands/install_uninstall_test.py

@@ -349,8 +349,9 @@ def test_install_existing_hooks_no_overwrite(tempdir_factory, store):
         # We should run both the legacy and pre-commit hooks
         ret, output = _get_commit_output(tempdir_factory)
         assert ret == 0
-        assert output.startswith('legacy hook\n')
-        NORMAL_PRE_COMMIT_RUN.assert_matches(output[len('legacy hook\n'):])
+        legacy = 'legacy hook\n'
+        assert output.startswith(legacy)
+        NORMAL_PRE_COMMIT_RUN.assert_matches(output.removeprefix(legacy))
 
 
 def test_legacy_overwriting_legacy_hook(tempdir_factory, store):
@@ -375,8 +376,9 @@ def test_install_existing_hook_no_overwrite_idempotent(tempdir_factory, store):
         # We should run both the legacy and pre-commit hooks
         ret, output = _get_commit_output(tempdir_factory)
         assert ret == 0
-        assert output.startswith('legacy hook\n')
-        NORMAL_PRE_COMMIT_RUN.assert_matches(output[len('legacy hook\n'):])
+        legacy = 'legacy hook\n'
+        assert output.startswith(legacy)
+        NORMAL_PRE_COMMIT_RUN.assert_matches(output.removeprefix(legacy))
 
 
 def test_install_with_existing_non_utf8_script(tmpdir, store):

tests/commands/migrate_config_test.py

@@ -1,10 +1,26 @@
 from __future__ import annotations
 
+from unittest import mock
+
 import pytest
+import yaml
 
 import pre_commit.constants as C
 from pre_commit.clientlib import InvalidConfigError
 from pre_commit.commands.migrate_config import migrate_config
+from pre_commit.yaml import yaml_compose
+
+
+@pytest.fixture(autouse=True, params=['c', 'pure'])
+def switch_pyyaml_impl(request):
+    if request.param == 'c':
+        yield
+    else:
+        with mock.patch.dict(
+                yaml_compose.keywords,
+                {'Loader': yaml.SafeLoader},
+        ):
+            yield
 
 
 def test_migrate_config_normal_format(tmpdir, capsys):
@@ -134,6 +150,27 @@ def test_migrate_config_sha_to_rev(tmpdir):
     )
 
 
+def test_migrate_config_sha_to_rev_json(tmp_path):
+    contents = """\
+{"repos": [{
+    "repo": "https://github.com/pre-commit/pre-commit-hooks",
+    "sha": "v1.2.0",
+    "hooks": []
+}]}
+"""
+    expected = """\
+{"repos": [{
+    "repo": "https://github.com/pre-commit/pre-commit-hooks",
+    "rev": "v1.2.0",
+    "hooks": []
+}]}
+"""
+    cfg = tmp_path.joinpath('cfg.yaml')
+    cfg.write_text(contents)
+    assert not migrate_config(str(cfg))
+    assert cfg.read_text() == expected
+
+
 def test_migrate_config_language_python_venv(tmp_path):
     src = '''\
 repos:
@@ -167,6 +204,73 @@ repos:
     assert cfg.read_text() == expected
 
 
+def test_migrate_config_quoted_python_venv(tmp_path):
+    src = '''\
+repos:
+-   repo: local
+    hooks:
+    -   id: example
+        name: example
+        entry: example
+        language: "python_venv"
+'''
+    expected = '''\
+repos:
+-   repo: local
+    hooks:
+    -   id: example
+        name: example
+        entry: example
+        language: "python"
+'''
+    cfg = tmp_path.joinpath('cfg.yaml')
+    cfg.write_text(src)
+    assert migrate_config(str(cfg)) == 0
+    assert cfg.read_text() == expected
+
+
+def test_migrate_config_default_stages(tmp_path):
+    src = '''\
+default_stages: [commit, push, merge-commit, commit-msg]
+repos: []
+'''
+    expected = '''\
+default_stages: [pre-commit, pre-push, pre-merge-commit, commit-msg]
+repos: []
+'''
+    cfg = tmp_path.joinpath('cfg.yaml')
+    cfg.write_text(src)
+    assert migrate_config(str(cfg)) == 0
+    assert cfg.read_text() == expected
+
+
+def test_migrate_config_hook_stages(tmp_path):
+    src = '''\
+repos:
+-   repo: local
+    hooks:
+    -   id: example
+        name: example
+        entry: example
+        language: system
+        stages: ["commit", "push", "merge-commit", "commit-msg"]
+'''
+    expected = '''\
+repos:
+-   repo: local
+    hooks:
+    -   id: example
+        name: example
+        entry: example
+        language: system
+        stages: ["pre-commit", "pre-push", "pre-merge-commit", "commit-msg"]
+'''
+    cfg = tmp_path.joinpath('cfg.yaml')
+    cfg.write_text(src)
+    assert migrate_config(str(cfg)) == 0
+    assert cfg.read_text() == expected
+
+
 def test_migrate_config_invalid_yaml(tmpdir):
     contents = '['
     cfg = tmpdir.join(C.CONFIG_FILE)

tests/commands/run_test.py

@@ -4,7 +4,7 @@ import os.path
 import shlex
 import sys
 import time
-from typing import MutableMapping
+from collections.abc import MutableMapping
 from unittest import mock
 
 import pytest
@@ -293,7 +293,7 @@ def test_verbose_duration(cap_out, store, in_git_dir, t1, t2, expected):
     write_config('.', {'repo': 'meta', 'hooks': [{'id': 'identity'}]})
     cmd_output('git', 'add', '.')
     opts = run_opts(verbose=True)
-    with mock.patch.object(time, 'time', side_effect=(t1, t2)):
+    with mock.patch.object(time, 'monotonic', side_effect=(t1, t2)):
         ret, printed = _do_run(cap_out, store, str(in_git_dir), opts)
     assert ret == 0
     assert expected in printed
@@ -1088,6 +1088,35 @@ def test_fail_fast_per_hook(cap_out, store, repo_with_failing_hook):
     assert printed.count(b'Failing hook') == 1
 
 
+def test_fail_fast_not_prev_failures(cap_out, store, repo_with_failing_hook):
+    with modify_config() as config:
+        config['repos'].append({
+            'repo': 'meta',
+            'hooks': [
+                {'id': 'identity', 'fail_fast': True},
+                {'id': 'identity', 'name': 'run me!'},
+            ],
+        })
+    stage_a_file()
+
+    ret, printed = _do_run(cap_out, store, repo_with_failing_hook, run_opts())
+    # should still run the last hook since the `fail_fast` one didn't fail
+    assert printed.count(b'run me!') == 1
+
+
+def test_fail_fast_run_arg(cap_out, store, repo_with_failing_hook):
+    with modify_config() as config:
+        # More than one hook to demonstrate early exit
+        config['repos'][0]['hooks'] *= 2
+    stage_a_file()
+
+    ret, printed = _do_run(
+        cap_out, store, repo_with_failing_hook, run_opts(fail_fast=True),
+    )
+    # it should have only run one hook due to the CLI flag
+    assert printed.count(b'Failing hook') == 1
+
+
 def test_classifier_removes_dne():
     classifier = Classifier(('this_file_does_not_exist',))
     assert classifier.filenames == []
@@ -1127,8 +1156,8 @@ def test_classifier_empty_types_or(tmpdir):
             types_or=[],
             exclude_types=[],
         )
-        assert for_symlink == ['foo']
-        assert for_file == ['bar']
+        assert tuple(for_symlink) == ('foo',)
+        assert tuple(for_file) == ('bar',)
 
 
 @pytest.fixture
@@ -1142,33 +1171,33 @@ def some_filenames():
 
 def test_include_exclude_base_case(some_filenames):
     ret = filter_by_include_exclude(some_filenames, '', '^$')
-    assert ret == [
+    assert tuple(ret) == (
         '.pre-commit-hooks.yaml',
         'pre_commit/git.py',
         'pre_commit/main.py',
-    ]
+    )
 
 
 def test_matches_broken_symlink(tmpdir):
     with tmpdir.as_cwd():
         os.symlink('does-not-exist', 'link')
         ret = filter_by_include_exclude({'link'}, '', '^$')
-        assert ret == ['link']
+        assert tuple(ret) == ('link',)
 
 
 def test_include_exclude_total_match(some_filenames):
     ret = filter_by_include_exclude(some_filenames, r'^.*\.py$', '^$')
-    assert ret == ['pre_commit/git.py', 'pre_commit/main.py']
+    assert tuple(ret) == ('pre_commit/git.py', 'pre_commit/main.py')
 
 
 def test_include_exclude_does_search_instead_of_match(some_filenames):
     ret = filter_by_include_exclude(some_filenames, r'\.yaml$', '^$')
-    assert ret == ['.pre-commit-hooks.yaml']
+    assert tuple(ret) == ('.pre-commit-hooks.yaml',)
 
 
 def test_include_exclude_exclude_removes_files(some_filenames):
     ret = filter_by_include_exclude(some_filenames, '', r'\.py$')
-    assert ret == ['.pre-commit-hooks.yaml']
+    assert tuple(ret) == ('.pre-commit-hooks.yaml',)
 
 
 def test_args_hook_only(cap_out, store, repo_with_passing_hook):

tests/commands/try_repo_test.py

@@ -43,7 +43,7 @@ def _run_try_repo(tempdir_factory, **kwargs):
 
 
 def test_try_repo_repo_only(cap_out, tempdir_factory):
-    with mock.patch.object(time, 'time', return_value=0.0):
+    with mock.patch.object(time, 'monotonic', return_value=0.0):
         _run_try_repo(tempdir_factory, verbose=True)
     start, config, rest = _get_out(cap_out)
     assert start == ''

tests/conftest.py

@@ -2,7 +2,6 @@ from __future__ import annotations
 
 import functools
 import io
-import logging
 import os.path
 from unittest import mock
 
@@ -203,42 +202,25 @@ def store(tempdir_factory):
     yield Store(os.path.join(tempdir_factory.get(), '.pre-commit'))
 
 
-@pytest.fixture
-def log_info_mock():
-    with mock.patch.object(logging.getLogger('pre_commit'), 'info') as mck:
-        yield mck
-
-
-class FakeStream:
-    def __init__(self):
-        self.data = io.BytesIO()
-
-    def write(self, s):
-        self.data.write(s)
-
-    def flush(self):
-        pass
-
-
 class Fixture:
-    def __init__(self, stream):
+    def __init__(self, stream: io.BytesIO) -> None:
         self._stream = stream
 
-    def get_bytes(self):
+    def get_bytes(self) -> bytes:
         """Get the output as-if no encoding occurred"""
-        data = self._stream.data.getvalue()
-        self._stream.data.seek(0)
-        self._stream.data.truncate()
+        data = self._stream.getvalue()
+        self._stream.seek(0)
+        self._stream.truncate()
         return data.replace(b'\r\n', b'\n')
 
-    def get(self):
+    def get(self) -> str:
         """Get the output assuming it was written as UTF-8 bytes"""
         return self.get_bytes().decode()
 
 
 @pytest.fixture
 def cap_out():
-    stream = FakeStream()
+    stream = io.BytesIO()
     write = functools.partial(output.write, stream=stream)
     write_line_b = functools.partial(output.write_line_b, stream=stream)
     with mock.patch.multiple(output, write=write, write_line_b=write_line_b):

tests/git_test.py

@@ -141,6 +141,15 @@ def test_get_conflicted_files_unstaged_files(in_merge_conflict):
     assert ret == {'conflict_file'}
 
 
+def test_get_conflicted_files_with_file_named_head(in_merge_conflict):
+    resolve_conflict()
+    open('HEAD', 'w').close()
+    cmd_output('git', 'add', 'HEAD')
+
+    ret = set(git.get_conflicted_files())
+    assert ret == {'conflict_file', 'HEAD'}
+
+
 MERGE_MSG = b"Merge branch 'foo' into bar\n\nConflicts:\n\tconflict_file\n"
 OTHER_MERGE_MSG = MERGE_MSG + b'\tother_conflict_file\n'
 

tests/lang_base_test.py

@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-import multiprocessing
 import os.path
 import sys
 from unittest import mock
@@ -10,6 +9,7 @@ import pytest
 import pre_commit.constants as C
 from pre_commit import lang_base
 from pre_commit import parse_shebang
+from pre_commit import xargs
 from pre_commit.prefix import Prefix
 from pre_commit.util import CalledProcessError
 
@@ -116,30 +116,23 @@ def test_no_env_noop(tmp_path):
     assert before == inside == after
 
 
-def test_target_concurrency_normal():
-    with mock.patch.object(multiprocessing, 'cpu_count', return_value=123):
-        with mock.patch.dict(os.environ, {}, clear=True):
-            assert lang_base.target_concurrency() == 123
+@pytest.fixture
+def cpu_count_mck():
+    with mock.patch.object(xargs, 'cpu_count', return_value=4):
+        yield
 
 
-def test_target_concurrency_testing_env_var():
-    with mock.patch.dict(
-            os.environ, {'PRE_COMMIT_NO_CONCURRENCY': '1'}, clear=True,
-    ):
-        assert lang_base.target_concurrency() == 1
-
-
-def test_target_concurrency_on_travis():
-    with mock.patch.dict(os.environ, {'TRAVIS': '1'}, clear=True):
-        assert lang_base.target_concurrency() == 2
-
-
-def test_target_concurrency_cpu_count_not_implemented():
-    with mock.patch.object(
-            multiprocessing, 'cpu_count', side_effect=NotImplementedError,
-    ):
-        with mock.patch.dict(os.environ, {}, clear=True):
-            assert lang_base.target_concurrency() == 1
+@pytest.mark.parametrize(
+    ('var', 'expected'),
+    (
+        ('PRE_COMMIT_NO_CONCURRENCY', 1),
+        ('TRAVIS', 2),
+        (None, 4),
+    ),
+)
+def test_target_concurrency(cpu_count_mck, var, expected):
+    with mock.patch.dict(os.environ, {var: '1'} if var else {}, clear=True):
+        assert lang_base.target_concurrency() == expected
 
 
 def test_shuffled_is_deterministic():
@@ -171,3 +164,15 @@ def test_basic_run_hook(tmp_path):
     assert ret == 0
     out = out.replace(b'\r\n', b'\n')
     assert out == b'hi hello file file file\n'
+
+
+def test_hook_cmd():
+    assert lang_base.hook_cmd('echo hi', ()) == ('echo', 'hi')
+
+
+def test_hook_cmd_hazmat():
+    ret = lang_base.hook_cmd('pre-commit hazmat cd a echo -- b', ())
+    assert ret == (
+        sys.executable, '-m', 'pre_commit.commands.hazmat',
+        'cd', 'a', 'echo', '--', 'b',
+    )

tests/languages/dart_test.py

@@ -10,7 +10,7 @@ from testing.language_helpers import run_language
 def test_dart(tmp_path):
     pubspec_yaml = '''\
 environment:
-  sdk: '>=2.10.0 <3.0.0'
+  sdk: '>=2.12.0 <4.0.0'
 
 name: hello_world_dart
 

tests/languages/docker_image_test.py

@@ -1,10 +1,18 @@
 from __future__ import annotations
 
+import pytest
+
 from pre_commit.languages import docker_image
+from pre_commit.util import cmd_output_b
 from testing.language_helpers import run_language
 from testing.util import xfailif_windows
 
 
+@pytest.fixture(autouse=True, scope='module')
+def _ensure_image_available():
+    cmd_output_b('docker', 'run', '--rm', 'ubuntu:22.04', 'echo')
+
+
 @xfailif_windows  # pragma: win32 no cover
 def test_docker_image_hook_via_entrypoint(tmp_path):
     ret = run_language(
@@ -25,3 +33,27 @@ def test_docker_image_hook_via_args(tmp_path):
         args=('hello hello world',),
     )
     assert ret == (0, b'hello hello world\n')
+
+
+@xfailif_windows  # pragma: win32 no cover
+def test_docker_image_color_tty(tmp_path):
+    ret = run_language(
+        tmp_path,
+        docker_image,
+        'ubuntu:22.04',
+        args=('grep', '--color', 'root', '/etc/group'),
+        color=True,
+    )
+    assert ret == (0, b'\x1b[01;31m\x1b[Kroot\x1b[m\x1b[K:x:0:\n')
+
+
+@xfailif_windows  # pragma: win32 no cover
+def test_docker_image_no_color_no_tty(tmp_path):
+    ret = run_language(
+        tmp_path,
+        docker_image,
+        'ubuntu:22.04',
+        args=('grep', '--color', 'root', '/etc/group'),
+        color=False,
+    )
+    assert ret == (0, b'root:x:0:\n')

tests/languages/docker_test.py

@@ -14,40 +14,173 @@ from pre_commit.util import CalledProcessError
 from testing.language_helpers import run_language
 from testing.util import xfailif_windows
 
-DOCKER_CGROUP_EXAMPLE = b'''\
-12:hugetlb:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-11:blkio:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-10:freezer:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-9:cpu,cpuacct:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-8:pids:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-7:rdma:/
-6:net_cls,net_prio:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-5:cpuset:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-4:devices:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-3:memory:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-2:perf_event:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-1:name=systemd:/docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7
-0::/system.slice/containerd.service
+DOCKER_CGROUPS_V1_MOUNTINFO_EXAMPLE = b'''\
+759 717 0:52 / / rw,relatime master:300 - overlay overlay rw,lowerdir=/var/lib/docker/overlay2/l/PCPE5P5IVGM7CFCPJR353N3ONK:/var/lib/docker/overlay2/l/EQFSDHFAJ333VEMEJD4ZTRIZCB,upperdir=/var/lib/docker/overlay2/0d9f6bf186030d796505b87d6daa92297355e47641e283d3c09d83a7f221e462/diff,workdir=/var/lib/docker/overlay2/0d9f6bf186030d796505b87d6daa92297355e47641e283d3c09d83a7f221e462/work
+760 759 0:58 / /proc rw,nosuid,nodev,noexec,relatime - proc proc rw
+761 759 0:59 / /dev rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+762 761 0:60 / /dev/pts rw,nosuid,noexec,relatime - devpts devpts rw,gid=5,mode=620,ptmxmode=666
+763 759 0:61 / /sys ro,nosuid,nodev,noexec,relatime - sysfs sysfs ro
+764 763 0:62 / /sys/fs/cgroup rw,nosuid,nodev,noexec,relatime - tmpfs tmpfs rw,mode=755,inode64
+765 764 0:29 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/systemd ro,nosuid,nodev,noexec,relatime master:11 - cgroup cgroup rw,xattr,name=systemd
+766 764 0:32 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/rdma ro,nosuid,nodev,noexec,relatime master:15 - cgroup cgroup rw,rdma
+767 764 0:33 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/cpu,cpuacct ro,nosuid,nodev,noexec,relatime master:16 - cgroup cgroup rw,cpu,cpuacct
+768 764 0:34 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/cpuset ro,nosuid,nodev,noexec,relatime master:17 - cgroup cgroup rw,cpuset
+769 764 0:35 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/pids ro,nosuid,nodev,noexec,relatime master:18 - cgroup cgroup rw,pids
+770 764 0:36 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/memory ro,nosuid,nodev,noexec,relatime master:19 - cgroup cgroup rw,memory
+771 764 0:37 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/perf_event ro,nosuid,nodev,noexec,relatime master:20 - cgroup cgroup rw,perf_event
+772 764 0:38 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/net_cls,net_prio ro,nosuid,nodev,noexec,relatime master:21 - cgroup cgroup rw,net_cls,net_prio
+773 764 0:39 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/blkio ro,nosuid,nodev,noexec,relatime master:22 - cgroup cgroup rw,blkio
+774 764 0:40 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/misc ro,nosuid,nodev,noexec,relatime master:23 - cgroup cgroup rw,misc
+775 764 0:41 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/hugetlb ro,nosuid,nodev,noexec,relatime master:24 - cgroup cgroup rw,hugetlb
+776 764 0:42 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/devices ro,nosuid,nodev,noexec,relatime master:25 - cgroup cgroup rw,devices
+777 764 0:43 /docker/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7 /sys/fs/cgroup/freezer ro,nosuid,nodev,noexec,relatime master:26 - cgroup cgroup rw,freezer
+778 761 0:57 / /dev/mqueue rw,nosuid,nodev,noexec,relatime - mqueue mqueue rw
+779 761 0:63 / /dev/shm rw,nosuid,nodev,noexec,relatime - tmpfs shm rw,size=65536k,inode64
+780 759 8:5 /var/lib/docker/containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/resolv.conf /etc/resolv.conf rw,relatime - ext4 /dev/sda5 rw,errors=remount-ro
+781 759 8:5 /var/lib/docker/containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/hostname /etc/hostname rw,relatime - ext4 /dev/sda5 rw,errors=remount-ro
+782 759 8:5 /var/lib/docker/containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/hosts /etc/hosts rw,relatime - ext4 /dev/sda5 rw,errors=remount-ro
+718 761 0:60 /0 /dev/console rw,nosuid,noexec,relatime - devpts devpts rw,gid=5,mode=620,ptmxmode=666
+719 760 0:58 /bus /proc/bus ro,nosuid,nodev,noexec,relatime - proc proc rw
+720 760 0:58 /fs /proc/fs ro,nosuid,nodev,noexec,relatime - proc proc rw
+721 760 0:58 /irq /proc/irq ro,nosuid,nodev,noexec,relatime - proc proc rw
+722 760 0:58 /sys /proc/sys ro,nosuid,nodev,noexec,relatime - proc proc rw
+723 760 0:58 /sysrq-trigger /proc/sysrq-trigger ro,nosuid,nodev,noexec,relatime - proc proc rw
+724 760 0:64 / /proc/asound ro,relatime - tmpfs tmpfs ro,inode64
+725 760 0:65 / /proc/acpi ro,relatime - tmpfs tmpfs ro,inode64
+726 760 0:59 /null /proc/kcore rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+727 760 0:59 /null /proc/keys rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+728 760 0:59 /null /proc/timer_list rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+729 760 0:66 / /proc/scsi ro,relatime - tmpfs tmpfs ro,inode64
+730 763 0:67 / /sys/firmware ro,relatime - tmpfs tmpfs ro,inode64
+731 763 0:68 / /sys/devices/virtual/powercap ro,relatime - tmpfs tmpfs ro,inode64
+'''  # noqa: E501
+
+DOCKER_CGROUPS_V2_MOUNTINFO_EXAMPLE = b'''\
+721 386 0:45 / / rw,relatime master:218 - overlay overlay rw,lowerdir=/var/lib/docker/overlay2/l/QHZ7OM7P4AQD3XLG274ZPWAJCV:/var/lib/docker/overlay2/l/5RFG6SZWVGOG2NKEYXJDQCQYX5,upperdir=/var/lib/docker/overlay2/e4ad859fc5d4791932b9b976052f01fb0063e01de3cef916e40ae2121f6a166e/diff,workdir=/var/lib/docker/overlay2/e4ad859fc5d4791932b9b976052f01fb0063e01de3cef916e40ae2121f6a166e/work,nouserxattr
+722 721 0:48 / /proc rw,nosuid,nodev,noexec,relatime - proc proc rw
+723 721 0:50 / /dev rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+724 723 0:51 / /dev/pts rw,nosuid,noexec,relatime - devpts devpts rw,gid=5,mode=620,ptmxmode=666
+725 721 0:52 / /sys ro,nosuid,nodev,noexec,relatime - sysfs sysfs ro
+726 725 0:26 / /sys/fs/cgroup ro,nosuid,nodev,noexec,relatime - cgroup2 cgroup rw,nsdelegate,memory_recursiveprot
+727 723 0:47 / /dev/mqueue rw,nosuid,nodev,noexec,relatime - mqueue mqueue rw
+728 723 0:53 / /dev/shm rw,nosuid,nodev,noexec,relatime - tmpfs shm rw,size=65536k,inode64
+729 721 8:3 /var/lib/docker/containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/resolv.conf /etc/resolv.conf rw,relatime - ext4 /dev/sda3 rw,errors=remount-ro
+730 721 8:3 /var/lib/docker/containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/hostname /etc/hostname rw,relatime - ext4 /dev/sda3 rw,errors=remount-ro
+731 721 8:3 /var/lib/docker/containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/hosts /etc/hosts rw,relatime - ext4 /dev/sda3 rw,errors=remount-ro
+387 723 0:51 /0 /dev/console rw,nosuid,noexec,relatime - devpts devpts rw,gid=5,mode=620,ptmxmode=666
+388 722 0:48 /bus /proc/bus ro,nosuid,nodev,noexec,relatime - proc proc rw
+389 722 0:48 /fs /proc/fs ro,nosuid,nodev,noexec,relatime - proc proc rw
+525 722 0:48 /irq /proc/irq ro,nosuid,nodev,noexec,relatime - proc proc rw
+526 722 0:48 /sys /proc/sys ro,nosuid,nodev,noexec,relatime - proc proc rw
+571 722 0:48 /sysrq-trigger /proc/sysrq-trigger ro,nosuid,nodev,noexec,relatime - proc proc rw
+572 722 0:57 / /proc/asound ro,relatime - tmpfs tmpfs ro,inode64
+575 722 0:58 / /proc/acpi ro,relatime - tmpfs tmpfs ro,inode64
+576 722 0:50 /null /proc/kcore rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+577 722 0:50 /null /proc/keys rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+578 722 0:50 /null /proc/timer_list rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,inode64
+579 722 0:59 / /proc/scsi ro,relatime - tmpfs tmpfs ro,inode64
+580 725 0:60 / /sys/firmware ro,relatime - tmpfs tmpfs ro,inode64
+'''  # noqa: E501
+
+PODMAN_CGROUPS_V1_MOUNTINFO_EXAMPLE = b'''\
+1200 915 0:57 / / rw,relatime - overlay overlay rw,lowerdir=/home/asottile/.local/share/containers/storage/overlay/l/ZWAU3VY3ZHABQJRBUAFPBX7R5D,upperdir=/home/asottile/.local/share/containers/storage/overlay/72504ef163fda63838930450553b7306412ccad139a007626732b3dc43af5200/diff,workdir=/home/asottile/.local/share/containers/storage/overlay/72504ef163fda63838930450553b7306412ccad139a007626732b3dc43af5200/work,volatile,userxattr
+1204 1200 0:62 / /proc rw,nosuid,nodev,noexec,relatime - proc proc rw
+1205 1200 0:63 / /dev rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,uid=1000,gid=1000,inode64
+1206 1200 0:64 / /sys ro,nosuid,nodev,noexec,relatime - sysfs sysfs rw
+1207 1205 0:65 / /dev/pts rw,nosuid,noexec,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
+1208 1205 0:61 / /dev/mqueue rw,nosuid,nodev,noexec,relatime - mqueue mqueue rw
+1209 1200 0:53 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/.containerenv /run/.containerenv rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=814036k,mode=700,uid=1000,gid=1000,inode64
+1210 1200 0:53 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/resolv.conf /etc/resolv.conf rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=814036k,mode=700,uid=1000,gid=1000,inode64
+1211 1200 0:53 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/hosts /etc/hosts rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=814036k,mode=700,uid=1000,gid=1000,inode64
+1212 1205 0:56 / /dev/shm rw,relatime - tmpfs shm rw,size=64000k,uid=1000,gid=1000,inode64
+1213 1200 0:53 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/hostname /etc/hostname rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=814036k,mode=700,uid=1000,gid=1000,inode64
+1214 1206 0:66 / /sys/fs/cgroup rw,nosuid,nodev,noexec,relatime - tmpfs cgroup rw,size=1024k,uid=1000,gid=1000,inode64
+1215 1214 0:43 / /sys/fs/cgroup/freezer ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,freezer
+1216 1214 0:42 /user.slice /sys/fs/cgroup/devices ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,devices
+1217 1214 0:41 / /sys/fs/cgroup/hugetlb ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,hugetlb
+1218 1214 0:40 / /sys/fs/cgroup/misc ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,misc
+1219 1214 0:39 / /sys/fs/cgroup/blkio ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,blkio
+1220 1214 0:38 / /sys/fs/cgroup/net_cls,net_prio ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,net_cls,net_prio
+1221 1214 0:37 / /sys/fs/cgroup/perf_event ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,perf_event
+1222 1214 0:36 /user.slice/user-1000.slice/user@1000.service /sys/fs/cgroup/memory ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,memory
+1223 1214 0:35 /user.slice/user-1000.slice/user@1000.service /sys/fs/cgroup/pids ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,pids
+1224 1214 0:34 / /sys/fs/cgroup/cpuset ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,cpuset
+1225 1214 0:33 / /sys/fs/cgroup/cpu,cpuacct ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,cpu,cpuacct
+1226 1214 0:32 / /sys/fs/cgroup/rdma ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,rdma
+1227 1214 0:29 /user.slice/user-1000.slice/user@1000.service/apps.slice/apps-org.gnome.Terminal.slice/vte-spawn-0c50448e-b395-4d76-8b92-379f16e5066f.scope /sys/fs/cgroup/systemd ro,nosuid,nodev,noexec,relatime - cgroup cgroup rw,xattr,name=systemd
+1228 1205 0:5 /null /dev/null rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1229 1205 0:5 /zero /dev/zero rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1230 1205 0:5 /full /dev/full rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1231 1205 0:5 /tty /dev/tty rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1232 1205 0:5 /random /dev/random rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1233 1205 0:5 /urandom /dev/urandom rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1234 1204 0:67 / /proc/acpi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+1235 1204 0:5 /null /proc/kcore rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1236 1204 0:5 /null /proc/keys rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1237 1204 0:5 /null /proc/timer_list rw,nosuid,noexec,relatime - devtmpfs udev rw,size=4031656k,nr_inodes=1007914,mode=755,inode64
+1238 1204 0:68 / /proc/scsi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+1239 1206 0:69 / /sys/firmware ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+1240 1206 0:70 / /sys/dev/block ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+1241 1204 0:62 /asound /proc/asound ro,relatime - proc proc rw
+1242 1204 0:62 /bus /proc/bus ro,relatime - proc proc rw
+1243 1204 0:62 /fs /proc/fs ro,relatime - proc proc rw
+1244 1204 0:62 /irq /proc/irq ro,relatime - proc proc rw
+1245 1204 0:62 /sys /proc/sys ro,relatime - proc proc rw
+1256 1204 0:62 /sysrq-trigger /proc/sysrq-trigger ro,relatime - proc proc rw
+916 1205 0:65 /0 /dev/console rw,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
+'''  # noqa: E501
+
+PODMAN_CGROUPS_V2_MOUNTINFO_EXAMPLE = b'''\
+685 690 0:63 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/resolv.conf /etc/resolv.conf rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=1637624k,nr_inodes=409406,mode=700,uid=1000,gid=1000,inode64
+686 690 0:63 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/hosts /etc/hosts rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=1637624k,nr_inodes=409406,mode=700,uid=1000,gid=1000,inode64
+687 692 0:50 / /dev/shm rw,nosuid,nodev,noexec,relatime - tmpfs shm rw,size=64000k,uid=1000,gid=1000,inode64
+688 690 0:63 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/.containerenv /run/.containerenv rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=1637624k,nr_inodes=409406,mode=700,uid=1000,gid=1000,inode64
+689 690 0:63 /containers/overlay-containers/c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7/userdata/hostname /etc/hostname rw,nosuid,nodev,relatime - tmpfs tmpfs rw,size=1637624k,nr_inodes=409406,mode=700,uid=1000,gid=1000,inode64
+690 546 0:55 / / rw,relatime - overlay overlay rw,lowerdir=/home/asottile/.local/share/containers/storage/overlay/l/NPOHYOD3PI3YW6TQSGBOVOUSK6,upperdir=/home/asottile/.local/share/containers/storage/overlay/565c206fb79f876ffd5f069b8bd7a97fb5e47d5d07396b0c395a4ed6725d4a8e/diff,workdir=/home/asottile/.local/share/containers/storage/overlay/565c206fb79f876ffd5f069b8bd7a97fb5e47d5d07396b0c395a4ed6725d4a8e/work,redirect_dir=nofollow,uuid=on,volatile,userxattr
+691 690 0:59 / /proc rw,nosuid,nodev,noexec,relatime - proc proc rw
+692 690 0:61 / /dev rw,nosuid - tmpfs tmpfs rw,size=65536k,mode=755,uid=1000,gid=1000,inode64
+693 690 0:62 / /sys ro,nosuid,nodev,noexec,relatime - sysfs sysfs rw
+694 692 0:66 / /dev/pts rw,nosuid,noexec,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
+695 692 0:58 / /dev/mqueue rw,nosuid,nodev,noexec,relatime - mqueue mqueue rw
+696 693 0:28 / /sys/fs/cgroup ro,nosuid,nodev,noexec,relatime - cgroup2 cgroup2 rw,nsdelegate,memory_recursiveprot
+698 692 0:6 /null /dev/null rw,nosuid,noexec,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+699 692 0:6 /zero /dev/zero rw,nosuid,noexec,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+700 692 0:6 /full /dev/full rw,nosuid,noexec,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+701 692 0:6 /tty /dev/tty rw,nosuid,noexec,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+702 692 0:6 /random /dev/random rw,nosuid,noexec,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+703 692 0:6 /urandom /dev/urandom rw,nosuid,noexec,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+704 691 0:67 / /proc/acpi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+705 691 0:6 /null /proc/kcore ro,nosuid,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+706 691 0:6 /null /proc/keys ro,nosuid,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+707 691 0:6 /null /proc/latency_stats ro,nosuid,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+708 691 0:6 /null /proc/timer_list ro,nosuid,relatime - devtmpfs udev rw,size=8147812k,nr_inodes=2036953,mode=755,inode64
+709 691 0:68 / /proc/scsi ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+710 693 0:69 / /sys/firmware ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+711 693 0:70 / /sys/dev/block ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+712 693 0:71 / /sys/devices/virtual/powercap ro,relatime - tmpfs tmpfs rw,size=0k,uid=1000,gid=1000,inode64
+713 691 0:59 /asound /proc/asound ro,nosuid,nodev,noexec,relatime - proc proc rw
+714 691 0:59 /bus /proc/bus ro,nosuid,nodev,noexec,relatime - proc proc rw
+715 691 0:59 /fs /proc/fs ro,nosuid,nodev,noexec,relatime - proc proc rw
+716 691 0:59 /irq /proc/irq ro,nosuid,nodev,noexec,relatime - proc proc rw
+717 691 0:59 /sys /proc/sys ro,nosuid,nodev,noexec,relatime - proc proc rw
+718 691 0:59 /sysrq-trigger /proc/sysrq-trigger ro,nosuid,nodev,noexec,relatime - proc proc rw
+547 692 0:66 /0 /dev/console rw,relatime - devpts devpts rw,gid=100004,mode=620,ptmxmode=666
 '''  # noqa: E501
 
 # The ID should match the above cgroup example.
 CONTAINER_ID = 'c33988ec7651ebc867cb24755eaf637a6734088bc7eef59d5799293a9e5450f7'  # noqa: E501
 
-NON_DOCKER_CGROUP_EXAMPLE = b'''\
-12:perf_event:/
-11:hugetlb:/
-10:devices:/
-9:blkio:/
-8:rdma:/
-7:cpuset:/
-6:cpu,cpuacct:/
-5:freezer:/
-4:memory:/
-3:pids:/
-2:net_cls,net_prio:/
-1:name=systemd:/init.scope
-0::/init.scope
-'''
+NON_DOCKER_MOUNTINFO_EXAMPLE = b'''\
+21 27 0:19 / /sys rw,nosuid,nodev,noexec,relatime shared:7 - sysfs sysfs rw
+22 27 0:20 / /proc rw,nosuid,nodev,noexec,relatime shared:14 - proc proc rw
+23 27 0:5 / /dev rw,nosuid,relatime shared:2 - devtmpfs udev rw,size=10219484k,nr_inodes=2554871,mode=755,inode64
+24 23 0:21 / /dev/pts rw,nosuid,noexec,relatime shared:3 - devpts devpts rw,gid=5,mode=620,ptmxmode=000
+25 27 0:22 / /run rw,nosuid,nodev,noexec,relatime shared:5 - tmpfs tmpfs rw,size=2047768k,mode=755,inode64
+27 1 8:2 / / rw,relatime shared:1 - ext4 /dev/sda2 rw,errors=remount-ro
+28 21 0:6 / /sys/kernel/security rw,nosuid,nodev,noexec,relatime shared:8 - securityfs securityfs rw
+29 23 0:24 / /dev/shm rw,nosuid,nodev shared:4 - tmpfs tmpfs rw,inode64
+30 25 0:25 / /run/lock rw,nosuid,nodev,noexec,relatime shared:6 - tmpfs tmpfs rw,size=5120k,inode64
+'''  # noqa: E501
 
 
 def test_docker_fallback_user():
@@ -62,9 +195,46 @@ def test_docker_fallback_user():
         assert docker.get_docker_user() == ()
 
 
-def test_in_docker_no_file():
+@pytest.fixture(autouse=True)
+def _avoid_cache():
+    with mock.patch.object(
+            docker,
+            '_is_rootless',
+            docker._is_rootless.__wrapped__,
+    ):
+        yield
+
+
+@pytest.mark.parametrize(
+    'info_ret',
+    (
+        (0, b'{"SecurityOptions": ["name=rootless","name=cgroupns"]}', b''),
+        (0, b'{"host": {"security": {"rootless": true}}}', b''),
+    ),
+)
+def test_docker_user_rootless(info_ret):
+    with mock.patch.object(docker, 'cmd_output_b', return_value=info_ret):
+        assert docker.get_docker_user() == ()
+
+
+@pytest.mark.parametrize(
+    'info_ret',
+    (
+        (0, b'{"SecurityOptions": ["name=cgroupns"]}', b''),
+        (0, b'{"host": {"security": {"rootless": false}}}', b''),
+        (0, b'{"response_from_some_other_container_engine": true}', b''),
+        (0, b'{"SecurityOptions": null}', b''),
+        (1, b'', b''),
+    ),
+)
+def test_docker_user_non_rootless(info_ret):
+    with mock.patch.object(docker, 'cmd_output_b', return_value=info_ret):
+        assert docker.get_docker_user() != ()
+
+
+def test_container_id_no_file():
     with mock.patch.object(builtins, 'open', side_effect=FileNotFoundError):
-        assert docker._is_in_docker() is False
+        assert docker._get_container_id() is None
 
 
 def _mock_open(data):
@@ -76,38 +246,33 @@ def _mock_open(data):
     )
 
 
-def test_in_docker_docker_in_file():
-    with _mock_open(DOCKER_CGROUP_EXAMPLE):
-        assert docker._is_in_docker() is True
-
-
-def test_in_docker_docker_not_in_file():
-    with _mock_open(NON_DOCKER_CGROUP_EXAMPLE):
-        assert docker._is_in_docker() is False
+def test_container_id_not_in_file():
+    with _mock_open(NON_DOCKER_MOUNTINFO_EXAMPLE):
+        assert docker._get_container_id() is None
 
 
 def test_get_container_id():
-    with _mock_open(DOCKER_CGROUP_EXAMPLE):
+    with _mock_open(DOCKER_CGROUPS_V1_MOUNTINFO_EXAMPLE):
+        assert docker._get_container_id() == CONTAINER_ID
+    with _mock_open(DOCKER_CGROUPS_V2_MOUNTINFO_EXAMPLE):
+        assert docker._get_container_id() == CONTAINER_ID
+    with _mock_open(PODMAN_CGROUPS_V1_MOUNTINFO_EXAMPLE):
+        assert docker._get_container_id() == CONTAINER_ID
+    with _mock_open(PODMAN_CGROUPS_V2_MOUNTINFO_EXAMPLE):
         assert docker._get_container_id() == CONTAINER_ID
 
 
-def test_get_container_id_failure():
-    with _mock_open(b''), pytest.raises(RuntimeError):
-        docker._get_container_id()
-
-
 def test_get_docker_path_not_in_docker_returns_same():
-    with mock.patch.object(docker, '_is_in_docker', return_value=False):
+    with _mock_open(b''):
         assert docker._get_docker_path('abc') == 'abc'
 
 
 @pytest.fixture
 def in_docker():
-    with mock.patch.object(docker, '_is_in_docker', return_value=True):
-        with mock.patch.object(
-            docker, '_get_container_id', return_value=CONTAINER_ID,
-        ):
-            yield
+    with mock.patch.object(
+        docker, '_get_container_id', return_value=CONTAINER_ID,
+    ):
+        yield
 
 
 def _linux_commonpath():
@@ -195,3 +360,14 @@ CMD ["echo", "This is overwritten by the entry"']
 
     ret = run_language(tmp_path, docker, 'echo hello hello world')
     assert ret == (0, b'hello hello world\n')
+
+
+@xfailif_windows  # pragma: win32 no cover
+def test_docker_hook_mount_permissions(tmp_path):
+    dockerfile = '''\
+FROM ubuntu:22.04
+'''
+    tmp_path.joinpath('Dockerfile').write_text(dockerfile)
+
+    retcode, _ = run_language(tmp_path, docker, 'touch', ('README.md',))
+    assert retcode == 0

tests/languages/dotnet_test.py

@@ -27,7 +27,7 @@ def _csproj(tool_name):
 <Project Sdk="Microsoft.NET.Sdk">
   <PropertyGroup>
     <OutputType>Exe</OutputType>
-    <TargetFramework>net6</TargetFramework>
+    <TargetFramework>net8</TargetFramework>
     <PackAsTool>true</PackAsTool>
     <ToolCommandName>{tool_name}</ToolCommandName>
     <PackageOutputPath>./nupkg</PackageOutputPath>

tests/languages/golang_test.py

@@ -7,10 +7,18 @@ import re_assert
 
 import pre_commit.constants as C
 from pre_commit import lang_base
+from pre_commit.commands.install_uninstall import install
 from pre_commit.envcontext import envcontext
 from pre_commit.languages import golang
 from pre_commit.store import _make_local_repo
+from pre_commit.util import CalledProcessError
+from pre_commit.util import cmd_output
+from testing.fixtures import add_config_to_repo
+from testing.fixtures import make_config_from_repo
 from testing.language_helpers import run_language
+from testing.util import cmd_output_mocked_pre_commit_home
+from testing.util import cwd
+from testing.util import git_commit
 
 
 ACTUAL_GET_DEFAULT_VERSION = golang.get_default_version.__wrapped__
@@ -111,11 +119,11 @@ def test_golang_versioned(tmp_path):
         tmp_path,
         golang,
         'go version',
-        version='1.18.4',
+        version='1.21.1',
     )
 
     assert ret == 0
-    assert out.startswith(b'go version go1.18.4')
+    assert out.startswith(b'go version go1.21.1')
 
 
 def test_local_golang_additional_deps(tmp_path):
@@ -128,9 +136,101 @@ def test_local_golang_additional_deps(tmp_path):
         deps=('golang.org/x/example/hello@latest',),
     )
 
-    assert ret == (0, b'Hello, Go examples!\n')
+    assert ret == (0, b'Hello, world!\n')
 
 
 def test_golang_hook_still_works_when_gobin_is_set(tmp_path):
     with envcontext((('GOBIN', str(tmp_path.joinpath('gobin'))),)):
         test_golang_system(tmp_path)
+
+
+def test_during_commit_all(tmp_path, tempdir_factory, store, in_git_dir):
+    hook_dir = tmp_path.joinpath('hook')
+    hook_dir.mkdir()
+    _make_hello_world(hook_dir)
+    hook_dir.joinpath('.pre-commit-hooks.yaml').write_text(
+        '-   id: hello-world\n'
+        '    name: hello world\n'
+        '    entry: golang-hello-world\n'
+        '    language: golang\n'
+        '    always_run: true\n',
+    )
+    cmd_output('git', 'init', hook_dir)
+    cmd_output('git', 'add', '.', cwd=hook_dir)
+    git_commit(cwd=hook_dir)
+
+    add_config_to_repo(in_git_dir, make_config_from_repo(hook_dir))
+
+    assert not install(C.CONFIG_FILE, store, hook_types=['pre-commit'])
+
+    git_commit(
+        fn=cmd_output_mocked_pre_commit_home,
+        tempdir_factory=tempdir_factory,
+    )
+
+
+def test_automatic_toolchain_switching(tmp_path):
+    go_mod = '''\
+module toolchain-version-test
+
+go 1.23.1
+'''
+    main_go = '''\
+package main
+
+func main() {}
+'''
+    tmp_path.joinpath('go.mod').write_text(go_mod)
+    mod_dir = tmp_path.joinpath('toolchain-version-test')
+    mod_dir.mkdir()
+    main_file = mod_dir.joinpath('main.go')
+    main_file.write_text(main_go)
+
+    with pytest.raises(CalledProcessError) as excinfo:
+        run_language(
+            path=tmp_path,
+            language=golang,
+            version='1.22.0',
+            exe='golang-version-test',
+        )
+
+    assert 'go.mod requires go >= 1.23.1' in excinfo.value.stderr.decode()
+
+
+def test_automatic_toolchain_switching_go_fmt(tmp_path, monkeypatch):
+    go_mod_hook = '''\
+module toolchain-version-test
+
+go 1.22.0
+'''
+    go_mod = '''\
+module toolchain-version-test
+
+go 1.23.1
+'''
+    main_go = '''\
+package main
+
+func main() {}
+'''
+    hook_dir = tmp_path.joinpath('hook')
+    hook_dir.mkdir()
+    hook_dir.joinpath('go.mod').write_text(go_mod_hook)
+
+    test_dir = tmp_path.joinpath('test')
+    test_dir.mkdir()
+    test_dir.joinpath('go.mod').write_text(go_mod)
+    main_file = test_dir.joinpath('main.go')
+    main_file.write_text(main_go)
+
+    with cwd(test_dir):
+        ret, out = run_language(
+            path=hook_dir,
+            language=golang,
+            version='1.22.0',
+            exe='go fmt',
+            file_args=(str(main_file),),
+        )
+
+        assert ret == 1
+        assert 'go.mod requires go >= 1.23.1' in out.decode()

tests/languages/haskell_test.py

@@ -0,0 +1,50 @@
+from __future__ import annotations
+
+import pytest
+
+from pre_commit.errors import FatalError
+from pre_commit.languages import haskell
+from pre_commit.util import win_exe
+from testing.language_helpers import run_language
+
+
+def test_run_example_executable(tmp_path):
+    example_cabal = '''\
+cabal-version:      2.4
+name:               example
+version:            0.1.0.0
+
+executable example
+    main-is:          Main.hs
+
+    build-depends:    base >=4
+    default-language: Haskell2010
+'''
+    main_hs = '''\
+module Main where
+
+main :: IO ()
+main = putStrLn "Hello, Haskell!"
+'''
+    tmp_path.joinpath('example.cabal').write_text(example_cabal)
+    tmp_path.joinpath('Main.hs').write_text(main_hs)
+
+    result = run_language(tmp_path, haskell, 'example')
+    assert result == (0, b'Hello, Haskell!\n')
+
+    # should not symlink things into environments
+    exe = tmp_path.joinpath(win_exe('hs_env-default/bin/example'))
+    assert exe.is_file()
+    assert not exe.is_symlink()
+
+
+def test_run_dep(tmp_path):
+    result = run_language(tmp_path, haskell, 'hello', deps=['hello'])
+    assert result == (0, b'Hello, World!\n')
+
+
+def test_run_empty(tmp_path):
+    with pytest.raises(FatalError) as excinfo:
+        run_language(tmp_path, haskell, 'example')
+    msg, = excinfo.value.args
+    assert msg == 'Expected .cabal files or additional_dependencies'

tests/languages/julia_test.py

@@ -0,0 +1,111 @@
+from __future__ import annotations
+
+import os
+from unittest import mock
+
+from pre_commit.languages import julia
+from testing.language_helpers import run_language
+from testing.util import cwd
+
+
+def _make_hook(tmp_path, julia_code):
+    src_dir = tmp_path.joinpath('src')
+    src_dir.mkdir()
+    src_dir.joinpath('main.jl').write_text(julia_code)
+    tmp_path.joinpath('Project.toml').write_text(
+        '[deps]\n'
+        'Example = "7876af07-990d-54b4-ab0e-23690620f79a"\n',
+    )
+
+
+def test_julia_hook(tmp_path):
+    code = """
+    using Example
+    function main()
+        println("Hello, world!")
+    end
+    main()
+    """
+    _make_hook(tmp_path, code)
+    expected = (0, b'Hello, world!\n')
+    assert run_language(tmp_path, julia, 'src/main.jl') == expected
+
+
+def test_julia_hook_with_startup(tmp_path):
+    depot_path = tmp_path.joinpath('depot')
+    depot_path.joinpath('config').mkdir(parents=True)
+    startup = depot_path.joinpath('config', 'startup.jl')
+    startup.write_text('error("Startup file used!")\n')
+
+    depo_path_var = f'{depot_path}{os.pathsep}'
+    with mock.patch.dict(os.environ, {'JULIA_DEPOT_PATH': depo_path_var}):
+        test_julia_hook(tmp_path)
+
+
+def test_julia_hook_manifest(tmp_path):
+    code = """
+    using Example
+    println(pkgversion(Example))
+    """
+    _make_hook(tmp_path, code)
+
+    tmp_path.joinpath('Manifest.toml').write_text(
+        'manifest_format = "2.0"\n\n'
+        '[[deps.Example]]\n'
+        'git-tree-sha1 = "11820aa9c229fd3833d4bd69e5e75ef4e7273bf1"\n'
+        'uuid = "7876af07-990d-54b4-ab0e-23690620f79a"\n'
+        'version = "0.5.4"\n',
+    )
+    expected = (0, b'0.5.4\n')
+    assert run_language(tmp_path, julia, 'src/main.jl') == expected
+
+
+def test_julia_hook_args(tmp_path):
+    code = """
+    function main(argv)
+        foreach(println, argv)
+    end
+    main(ARGS)
+    """
+    _make_hook(tmp_path, code)
+    expected = (0, b'--arg1\n--arg2\n')
+    assert run_language(
+        tmp_path, julia, 'src/main.jl --arg1 --arg2',
+    ) == expected
+
+
+def test_julia_hook_additional_deps(tmp_path):
+    code = """
+    using TOML
+    function main()
+        project_file = Base.active_project()
+        dict = TOML.parsefile(project_file)
+        for (k, v) in dict["deps"]
+            println(k, " = ", v)
+        end
+    end
+    main()
+    """
+    _make_hook(tmp_path, code)
+    deps = ('TOML=fa267f1f-6049-4f14-aa54-33bafae1ed76',)
+    ret, out = run_language(tmp_path, julia, 'src/main.jl', deps=deps)
+    assert ret == 0
+    assert b'Example = 7876af07-990d-54b4-ab0e-23690620f79a' in out
+    assert b'TOML = fa267f1f-6049-4f14-aa54-33bafae1ed76' in out
+
+
+def test_julia_repo_local(tmp_path):
+    env_dir = tmp_path.joinpath('envdir')
+    env_dir.mkdir()
+    local_dir = tmp_path.joinpath('local')
+    local_dir.mkdir()
+    local_dir.joinpath('local.jl').write_text(
+        'using TOML; foreach(println, ARGS)',
+    )
+    with cwd(local_dir):
+        deps = ('TOML=fa267f1f-6049-4f14-aa54-33bafae1ed76',)
+        expected = (0, b'--local-arg1\n--local-arg2\n')
+        assert run_language(
+            env_dir, julia, 'local.jl --local-arg1 --local-arg2',
+            deps=deps, is_local=True,
+        ) == expected

tests/languages/node_test.py

@@ -139,7 +139,7 @@ def test_node_with_user_config_set(tmp_path):
         test_node_hook_system(tmp_path)
 
 
-@pytest.mark.parametrize('version', (C.DEFAULT, '18.13.0'))
+@pytest.mark.parametrize('version', (C.DEFAULT, '18.14.0'))
 def test_node_hook_versions(tmp_path, version):
     _make_hello_world(tmp_path)
     ret = run_language(tmp_path, node, 'node-hello', version=version)

tests/languages/python_test.py

@@ -10,8 +10,11 @@ import pre_commit.constants as C
 from pre_commit.envcontext import envcontext
 from pre_commit.languages import python
 from pre_commit.prefix import Prefix
+from pre_commit.store import _make_local_repo
+from pre_commit.util import cmd_output_b
 from pre_commit.util import make_executable
 from pre_commit.util import win_exe
+from testing.auto_namedtuple import auto_namedtuple
 from testing.language_helpers import run_language
 
 
@@ -34,6 +37,72 @@ def test_read_pyvenv_cfg_non_utf8(tmpdir):
     assert python._read_pyvenv_cfg(pyvenv_cfg) == expected
 
 
+def _get_default_version(
+        *,
+        impl: str,
+        exe: str,
+        found: set[str],
+        version: tuple[int, int],
+) -> str:
+    sys_exe = f'/fake/path/{exe}'
+    sys_impl = auto_namedtuple(name=impl)
+    sys_ver = auto_namedtuple(major=version[0], minor=version[1])
+
+    def find_exe(s):
+        if s in found:
+            return f'/fake/path/found/{exe}'
+        else:
+            return None
+
+    with (
+            mock.patch.object(sys, 'implementation', sys_impl),
+            mock.patch.object(sys, 'executable', sys_exe),
+            mock.patch.object(sys, 'version_info', sys_ver),
+            mock.patch.object(python, 'find_executable', find_exe),
+    ):
+        return python.get_default_version.__wrapped__()
+
+
+def test_default_version_sys_executable_found():
+    ret = _get_default_version(
+        impl='cpython',
+        exe='python3.12',
+        found={'python3.12'},
+        version=(3, 12),
+    )
+    assert ret == 'python3.12'
+
+
+def test_default_version_picks_specific_when_found():
+    ret = _get_default_version(
+        impl='cpython',
+        exe='python3',
+        found={'python3', 'python3.12'},
+        version=(3, 12),
+    )
+    assert ret == 'python3.12'
+
+
+def test_default_version_picks_pypy_versioned_exe():
+    ret = _get_default_version(
+        impl='pypy',
+        exe='python',
+        found={'pypy3.12', 'python3'},
+        version=(3, 12),
+    )
+    assert ret == 'pypy3.12'
+
+
+def test_default_version_picks_pypy_unversioned_exe():
+    ret = _get_default_version(
+        impl='pypy',
+        exe='python',
+        found={'pypy3', 'python3'},
+        version=(3, 12),
+    )
+    assert ret == 'pypy3'
+
+
 def test_norm_version_expanduser():
     home = os.path.expanduser('~')
     if sys.platform == 'win32':  # pragma: win32 cover
@@ -284,3 +353,15 @@ def test_python_hook_weird_setup_cfg(tmp_path):
 
     ret = run_language(tmp_path, python, 'socks', [os.devnull])
     assert ret == (0, f'[{os.devnull!r}]\nhello hello\n'.encode())
+
+
+def test_local_repo_with_other_artifacts(tmp_path):
+    cmd_output_b('git', 'init', tmp_path)
+    _make_local_repo(str(tmp_path))
+    # pretend a rust install also ran here
+    tmp_path.joinpath('target').mkdir()
+
+    ret, out = run_language(tmp_path, python, 'python --version')
+
+    assert ret == 0
+    assert out.startswith(b'Python ')

tests/languages/r_test.py

@@ -1,14 +1,17 @@
 from __future__ import annotations
 
 import os.path
-import shutil
+from unittest import mock
 
 import pytest
 
+import pre_commit.constants as C
 from pre_commit import envcontext
+from pre_commit import lang_base
 from pre_commit.languages import r
 from pre_commit.prefix import Prefix
 from pre_commit.store import _make_local_repo
+from pre_commit.util import resource_text
 from pre_commit.util import win_exe
 from testing.language_helpers import run_language
 
@@ -127,7 +130,8 @@ def test_path_rscript_exec_no_r_home_set():
         assert r._rscript_exec() == 'Rscript'
 
 
-def test_r_hook(tmp_path):
+@pytest.fixture
+def renv_lock_file(tmp_path):
     renv_lock = '''\
 {
   "R": {
@@ -157,6 +161,12 @@ def test_r_hook(tmp_path):
   }
 }
 '''
+    tmp_path.joinpath('renv.lock').write_text(renv_lock)
+    yield
+
+
+@pytest.fixture
+def description_file(tmp_path):
     description = '''\
 Package: gli.clu
 Title: What the Package Does (One Line, Title Case)
@@ -178,27 +188,39 @@ RoxygenNote: 7.1.1
 Imports:
     rprojroot
 '''
-    hello_world_r = '''\
+    tmp_path.joinpath('DESCRIPTION').write_text(description)
+    yield
+
+
+@pytest.fixture
+def hello_world_file(tmp_path):
+    hello_world = '''\
 stopifnot(
     packageVersion('rprojroot') == '1.0',
     packageVersion('gli.clu') == '0.0.0.9000'
 )
 cat("Hello, World, from R!\n")
 '''
+    tmp_path.joinpath('hello-world.R').write_text(hello_world)
+    yield
 
-    tmp_path.joinpath('renv.lock').write_text(renv_lock)
-    tmp_path.joinpath('DESCRIPTION').write_text(description)
-    tmp_path.joinpath('hello-world.R').write_text(hello_world_r)
+
+@pytest.fixture
+def renv_folder(tmp_path):
     renv_dir = tmp_path.joinpath('renv')
     renv_dir.mkdir()
-    shutil.copy(
-        os.path.join(
-            os.path.dirname(__file__),
-            '../../pre_commit/resources/empty_template_activate.R',
-        ),
-        renv_dir.joinpath('activate.R'),
-    )
+    activate_r = resource_text('empty_template_activate.R')
+    renv_dir.joinpath('activate.R').write_text(activate_r)
+    yield
 
+
+def test_r_hook(
+        tmp_path,
+        renv_lock_file,
+        description_file,
+        hello_world_file,
+        renv_folder,
+):
     expected = (0, b'Hello, World, from R!\n')
     assert run_language(tmp_path, r, 'Rscript hello-world.R') == expected
 
@@ -221,3 +243,55 @@ Rscript -e '
         args=('hi', 'hello'),
     )
     assert ret == (0, b'hi, hello, from R!\n')
+
+
+@pytest.fixture
+def prefix(tmpdir):
+    yield Prefix(str(tmpdir))
+
+
+@pytest.fixture
+def installed_environment(
+        renv_lock_file,
+        hello_world_file,
+        renv_folder,
+        prefix,
+):
+    env_dir = lang_base.environment_dir(
+        prefix, r.ENVIRONMENT_DIR, r.get_default_version(),
+    )
+    r.install_environment(prefix, C.DEFAULT, ())
+    yield prefix, env_dir
+
+
+def test_health_check_healthy(installed_environment):
+    # should be healthy right after creation
+    prefix, _ = installed_environment
+    assert r.health_check(prefix, C.DEFAULT) is None
+
+
+def test_health_check_after_downgrade(installed_environment):
+    prefix, _ = installed_environment
+
+    # pretend the saved installed version is old
+    with mock.patch.object(r, '_read_installed_version', return_value='1.0.0'):
+        output = r.health_check(prefix, C.DEFAULT)
+
+    assert output is not None
+    assert output.startswith('Hooks were installed for R version')
+
+
+@pytest.mark.parametrize('version', ('NULL', 'NA', "''"))
+def test_health_check_without_version(prefix, installed_environment, version):
+    prefix, env_dir = installed_environment
+
+    # simulate old pre-commit install by unsetting the installed version
+    r._execute_r_in_renv(
+        f'renv::settings$r.version({version})',
+        prefix=prefix, version=C.DEFAULT, cwd=env_dir,
+    )
+
+    # no R version specified fails as unhealty
+    msg = 'Hooks were installed with an unknown R version'
+    check_output = r.health_check(prefix, C.DEFAULT)
+    assert check_output is not None and check_output.startswith(msg)

tests/languages/ruby_test.py

@@ -91,8 +91,8 @@ def test_ruby_additional_deps(tmp_path):
         tmp_path,
         ruby,
         'ruby -e',
-        args=('require "tins"',),
-        deps=('tins',),
+        args=('require "jmespath"',),
+        deps=('jmespath',),
     )
     assert ret == (0, b'')
 

tests/languages/rust_test.py

@@ -9,6 +9,7 @@ from pre_commit import parse_shebang
 from pre_commit.languages import rust
 from pre_commit.store import _make_local_repo
 from testing.language_helpers import run_language
+from testing.util import cwd
 
 ACTUAL_GET_DEFAULT_VERSION = rust.get_default_version.__wrapped__
 
@@ -29,6 +30,14 @@ def test_uses_default_when_rust_is_not_available(cmd_output_b_mck):
     assert ACTUAL_GET_DEFAULT_VERSION() == C.DEFAULT
 
 
+def test_selects_system_even_if_rust_toolchain_toml(tmp_path):
+    toolchain_toml = '[toolchain]\nchannel = "wtf"\n'
+    tmp_path.joinpath('rust-toolchain.toml').write_text(toolchain_toml)
+
+    with cwd(tmp_path):
+        assert ACTUAL_GET_DEFAULT_VERSION() == 'system'
+
+
 def _make_hello_world(tmp_path):
     src_dir = tmp_path.joinpath('src')
     src_dir.mkdir()

tests/languages/system_test.py

@@ -1,9 +0,0 @@
-from __future__ import annotations
-
-from pre_commit.languages import system
-from testing.language_helpers import run_language
-
-
-def test_system_language(tmp_path):
-    expected = (0, b'hello hello world\n')
-    assert run_language(tmp_path, system, 'echo hello hello world') == expected

tests/languages/unsupported_script_test.py

@@ -1,14 +1,14 @@
 from __future__ import annotations
 
-from pre_commit.languages import script
+from pre_commit.languages import unsupported_script
 from pre_commit.util import make_executable
 from testing.language_helpers import run_language
 
 
-def test_script_language(tmp_path):
+def test_unsupported_script_language(tmp_path):
     exe = tmp_path.joinpath('main')
     exe.write_text('#!/usr/bin/env bash\necho hello hello world\n')
     make_executable(exe)
 
     expected = (0, b'hello hello world\n')
-    assert run_language(tmp_path, script, 'main') == expected
+    assert run_language(tmp_path, unsupported_script, 'main') == expected