fix: ignore gitignore patterns with unmatched trailing backslashes

CHANGES_1.in.rst

@@ -16,6 +16,7 @@ New features:
 
 Bug fixes:
 
+- Ignore gitignore patterns ending with an unmatched backslash.
 - `Pull #123`_: Ignore invalid gitignore bracket ranges for `GitIgnoreSpec`.
 - `Pull #128`_: Support POSIX character classes (e.g. `[[:alpha:]]`) in gitignore bracket expressions.
 - `Issue #129`_ / `Pull #132`_: Fix GitIgnoreSpec re-including files under an excluded directory

pathspec/patterns/gitignore/basic.py

@@ -179,6 +179,14 @@ def pattern_to_regex(
 			# a literal hash (i.e., '\#').
 			return (None, None)
 
+		elif original_pattern.rstrip('\r\n').endswith('\\'):
+			pattern_line = original_pattern.rstrip('\r\n')
+			trailing_backslashes = len(pattern_line) - len(pattern_line.rstrip('\\'))
+			if trailing_backslashes % 2:
+				# A pattern ending with an unmatched backslash is invalid and never
+				# matches.
+				return (None, None)
+
 		if pattern_str.startswith('!'):
 			# A pattern starting with an exclamation mark ('!') negates the pattern
 			# (exclude instead of include). Escape the exclamation mark with a back

pathspec/patterns/gitignore/spec.py

@@ -213,6 +213,14 @@ def pattern_to_regex(
 			# a literal hash (i.e., '\#').
 			return (None, None)
 
+		elif original_pattern.rstrip('\r\n').endswith('\\'):
+			pattern_line = original_pattern.rstrip('\r\n')
+			trailing_backslashes = len(pattern_line) - len(pattern_line.rstrip('\\'))
+			if trailing_backslashes % 2:
+				# A pattern ending with an unmatched backslash is invalid and never
+				# matches.
+				return (None, None)
+
 		elif pattern_str == '/':
 			# EDGE CASE: According to `git check-ignore` (v2.4.1), a single '/' does
 			# not match any file.

tests/test_03_gitignore_basic.py

@@ -616,11 +616,15 @@ def test_08_escape(self):
 		result = GitIgnoreBasicPattern.escape(fname)
 		self.assertEqual(result, escaped)
 
-	def test_09_single_escape_fail(self):
+	def test_09_single_backslash(self):
 		"""
-		Test an escape on a line by itself.
+		Test that a lone backslash is an invalid pattern that never matches. Git
+		treats invalid gitignore patterns as null patterns instead of rejecting the
+		entire set of patterns.
 		"""
-		self._check_invalid_pattern('\\')
+		pattern = GitIgnoreBasicPattern('\\')
+		self.assertIs(pattern.include, None)
+		self.assertIs(pattern.regex, None)
 
 	def test_09_single_exclamation_mark_fail(self):
 		"""
@@ -984,6 +988,21 @@ def test_16_repr_str(self):
 		self.assertEqual(repr(pattern), "GitIgnoreBasicPattern(pattern='*.py', include=True)")
 		self.assertEqual(str(pattern), '*.py')
 
+	def test_17_trailing_backslash(self):
+		"""
+		Test that a pattern ending in one unmatched backslash is invalid, while
+		two trailing backslashes encode one literal backslash.
+		"""
+		for raw_pattern in ['fileA\\', 'fileA\\\n']:
+			with self.subTest(f"p={raw_pattern!r}"):
+				pattern = GitIgnoreBasicPattern(raw_pattern)
+				self.assertIs(pattern.include, None)
+				self.assertIs(pattern.regex, None)
+
+		pattern = GitIgnoreBasicPattern('fileA\\\\')
+		self.assertIs(pattern.include, True)
+		self.assertIsNotNone(pattern.match_file('fileA\\'))
+
 	def test_globstars_match_newlines(self):
 		for pattern, path in [
 			("target", "line\nbreak/target"),

tests/test_04_gitignore_spec.py

@@ -634,11 +634,15 @@ def test_08_escape_trailing_space(self):
 			pattern = GitIgnoreSpecPattern(escaped)
 			self.assertEqual(set(filter(pattern.match_file, [fname])), {fname}, (fname, escaped))
 
-	def test_09_single_escape_fail(self):
+	def test_09_single_backslash(self):
 		"""
-		Test an escape on a line by itself.
+		Test that a lone backslash is an invalid pattern that never matches. Git
+		treats invalid gitignore patterns as null patterns instead of rejecting the
+		entire set of patterns.
 		"""
-		self._check_invalid_pattern('\\')
+		pattern = GitIgnoreSpecPattern('\\')
+		self.assertIs(pattern.include, None)
+		self.assertIs(pattern.regex, None)
 
 	def test_09_single_exclamation_mark_fail(self):
 		"""
@@ -1111,3 +1115,18 @@ def test_16_posix_class_e_invalid(self):
 				pattern = GitIgnoreSpecPattern(raw_pattern)
 				self.assertIs(pattern.include, None)
 				self.assertIs(pattern.regex, None)
+
+	def test_17_trailing_backslash(self):
+		"""
+		Test that a pattern ending in one unmatched backslash is invalid, while
+		two trailing backslashes encode one literal backslash.
+		"""
+		for raw_pattern in ['fileA\\', 'fileA\\\n']:
+			with self.subTest(f"p={raw_pattern!r}"):
+				pattern = GitIgnoreSpecPattern(raw_pattern)
+				self.assertIs(pattern.include, None)
+				self.assertIs(pattern.regex, None)
+
+		pattern = GitIgnoreSpecPattern('fileA\\\\')
+		self.assertIs(pattern.include, True)
+		self.assertIsNotNone(pattern.match_file('fileA\\'))

tests/test_05_pathspec.py

@@ -1132,3 +1132,20 @@ def test_11_repr(self):
 			repr(spec),
 			"PathSpec(patterns=[GitIgnoreBasicPattern(pattern='*.py', include=True)], backend='simple')",
 		)
+
+	def test_12_trailing_backslash(self):
+		"""
+		Test that an invalid trailing-backslash pattern does not prevent a later
+		valid pattern from being used.
+		"""
+		spec = PathSpec.from_lines('gitignore', [
+			'fileA\\',
+			'*.log',
+		], backend='simple')
+
+		self.assertIs(spec.patterns[0].include, None)
+		self.assertIs(spec.patterns[0].regex, None)
+		self.assertIs(spec.match_file('fileA'), False)
+		self.assertIs(spec.match_file('fileA\\'), False)
+		self.assertIs(spec.match_file('fileA.log'), True)
+		self.assertIs(spec.match_file('fileA.txt'), False)

tests/test_06_gitignore.py

@@ -955,3 +955,20 @@ def test_13_issue_139(self):
 			for sub_test in self.parameterize_from_lines([pattern]):
 				with sub_test() as spec:
 					self.assertTrue(spec.match_file(path))
+
+	def test_14_trailing_backslash(self):
+		"""
+		Test that an invalid trailing-backslash pattern does not prevent a later
+		valid pattern from being used.
+		"""
+		spec = GitIgnoreSpec.from_lines([
+			'fileA\\',
+			'*.log',
+		], backend='simple')
+
+		self.assertIs(spec.patterns[0].include, None)
+		self.assertIs(spec.patterns[0].regex, None)
+		self.assertIs(spec.match_file('fileA'), False)
+		self.assertIs(spec.match_file('fileA\\'), False)
+		self.assertIs(spec.match_file('fileA.log'), True)
+		self.assertIs(spec.match_file('fileA.txt'), False)