Add check for permitted characters in pathnames

Add a new checker that checks for allowed characters in pathnames.

Includes arguments for overriding the default allowed list and extending
it.

Default allow list is alphanumeric plus `.-_`.

.pre-commit-hooks.yaml

@@ -45,6 +45,12 @@
     entry: check-json
     language: python
     types: [json]
+-   id: check-permitted-path-characters
+    name: check for permitted path characters
+    description: checks pathnames for allowed characters
+    entry: check-permitted-path-characters
+    language: python
+    args: ['--']
 -   id: check-shebang-scripts-are-executable
     name: check that scripts with shebangs are executable
     description: ensures that (non-binary) files with a shebang are executable.

README.md

@@ -58,6 +58,9 @@ Attempts to load all json files to verify syntax.
 #### `check-merge-conflict`
 Check for files that contain merge conflict strings.
 
+#### `check-permitted-path-characters`
+Check for allowed characters in pathnames.
+
 #### `check-shebang-scripts-are-executable`
 Checks that scripts with shebangs are executable.
 

pre_commit_hooks/check_permitted_path_characters.py

@@ -0,0 +1,49 @@
+from __future__ import annotations
+
+import argparse
+import string
+from typing import Sequence
+
+DEFAULT_ALLOWLIST = string.ascii_letters + string.digits + '-_./'
+
+
+def main(argv: Sequence[str] | None = None) -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument('filenames', nargs='*')
+    parser.add_argument(
+        '--allowlist',
+        default=DEFAULT_ALLOWLIST,
+        help=(
+            'Override the default allowlist of permitted characters. The'
+            ' default is %(default)s'
+        ),
+    )
+    parser.add_argument(
+        '--extra-allowlist',
+        default='',
+        help='Extend the default allowlist with these characters.',
+    )
+    args = parser.parse_args(argv)
+
+    allowlist = set(args.allowlist + args.extra_allowlist)
+
+    retcode = 0
+    for filename in args.filenames:
+        # check the entire path, not just the filename, to catch directories
+        # with invalid characters
+        file_chars = set(filename)
+        if not file_chars.issubset(allowlist):
+            # sorted and stringified for readability
+            pretty_allowlist = ''.join(sorted(allowlist))
+            pretty_banlist = repr(''.join(sorted(file_chars - allowlist)))
+            print(
+                f'"{filename}" contains characters not in the allowlist:'
+                f' "{pretty_banlist}". Allowlist is: "{pretty_allowlist}".',
+            )
+            retcode = 1
+
+    return retcode
+
+
+if __name__ == '__main__':
+    raise SystemExit(main())

setup.cfg

@@ -43,6 +43,7 @@ console_scripts =
     check-executables-have-shebangs = pre_commit_hooks.check_executables_have_shebangs:main
     check-json = pre_commit_hooks.check_json:main
     check-merge-conflict = pre_commit_hooks.check_merge_conflict:main
+    check-permitted-path-characters = pre_commit_hooks.check_permitted_path_characters:main
     check-shebang-scripts-are-executable = pre_commit_hooks.check_shebang_scripts_are_executable:main
     check-symlinks = pre_commit_hooks.check_symlinks:main
     check-toml = pre_commit_hooks.check_toml:main

tests/check_permitted_path_characters_test.py

@@ -0,0 +1,50 @@
+from __future__ import annotations
+
+from pre_commit_hooks.check_permitted_path_characters import main
+
+
+def test_main_all_pass():
+    ret = main(
+        [
+            '/some/path/foo_test.py',
+            './relative/path/bar_test.py',
+            'filename-only.py',
+        ],
+    )
+    assert ret == 0
+
+
+def test_main_default_chars():
+    # use '--' for separating pathnames from args, so pathnames with leading
+    # '-' are not interpreted as flags
+    ret = main(
+        [
+            '--',
+            '-.0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_',
+            'abcdefghijklmnopqrstuvwxyz',
+        ],
+    )
+    assert ret == 0
+
+    ret = main(['+'])
+    assert ret == 1
+
+
+def test_main_invalid_dir():
+    ret = main(['--', '/some+funky%%dir/pathname'])
+    assert ret == 1
+
+
+def test_main_allowlist():
+    ret = main(['--allowlist', 'abc', 'invalid.py'])
+    assert ret == 1
+    ret = main(['--allowlist', 'abc', 'cba'])
+    assert ret == 0
+    # a pathological case
+    ret = main(['--allowlist', '\b\x01\t/.', '\b.\x01/\t'])
+    assert ret == 0
+
+
+def test_main_extra_allowlist():
+    ret = main(['--extra-allowlist', '+', 'valid+'])
+    assert ret == 0