Ignore quantifiers when splitting comma-separated regexes

doc/whatsnew/fragments/7229.bugfix

@@ -0,0 +1,5 @@
+When parsing comma-separated lists of regular expressions in the config, ignore
+commas that are inside braces since those indicate quantiers, not dilineation
+between expressions.
+
+Closes #7229

pylint/config/argument.py

@@ -102,7 +102,7 @@ def _py_version_transformer(value: str) -> tuple[int, ...]:
 def _regexp_csv_transfomer(value: str) -> Sequence[Pattern[str]]:
     """Transforms a comma separated list of regular expressions."""
     patterns: list[Pattern[str]] = []
-    for pattern in _csv_transformer(value):
+    for pattern in pylint_utils._check_regexp_csv(value):
         patterns.append(re.compile(pattern))
     return patterns
 

pylint/utils/__init__.py

@@ -14,6 +14,7 @@
     HAS_ISORT_5,
     IsortDriver,
     _check_csv,
+    _check_regexp_csv,
     _format_option_value,
     _splitstrip,
     _unquote,
@@ -34,6 +35,7 @@
     "HAS_ISORT_5",
     "IsortDriver",
     "_check_csv",
+    "_check_regexp_csv",
     "_format_option_value",
     "_splitstrip",
     "_unquote",

pylint/utils/utils.py

@@ -21,7 +21,8 @@
 import textwrap
 import tokenize
 import warnings
-from collections.abc import Sequence
+from collections import deque
+from collections.abc import Iterable, Sequence
 from io import BufferedReader, BytesIO
 from typing import (
     TYPE_CHECKING,
@@ -328,6 +329,29 @@ def _check_csv(value: list[str] | tuple[str] | str) -> Sequence[str]:
     return _splitstrip(value)
 
 
+def _check_regexp_csv(value: list[str] | tuple[str] | str) -> Iterable[str]:
+    if isinstance(value, (list, tuple)):
+        yield from value
+    else:
+        # None is a sentinel value here
+        regexps: deque[deque[str] | None] = deque([None])
+        open_braces = False
+        for char in value:
+            if char == "{":
+                open_braces = True
+            elif char == "}" and open_braces:
+                open_braces = False
+
+            if char == "," and not open_braces:
+                regexps.append(None)
+            elif regexps[-1] is None:
+                regexps.pop()
+                regexps.append(deque([char]))
+            else:
+                regexps[-1].append(char)
+        yield from ("".join(regexp).strip() for regexp in regexps if regexp is not None)
+
+
 def _comment(string: str) -> str:
     """Return string as a comment."""
     lines = [line.strip() for line in string.splitlines()]

tests/config/test_config.py

@@ -5,6 +5,8 @@
 from __future__ import annotations
 
 import os
+import re
+import timeit
 from pathlib import Path
 
 import pytest
@@ -111,6 +113,31 @@ def test_unknown_py_version(capsys: CaptureFixture) -> None:
     assert "the-newest has an invalid format, should be a version string." in output.err
 
 
+CSV_REGEX_COMMA_CASES = [
+    ("foo", ["foo"]),
+    ("foo,bar", ["foo", "bar"]),
+    ("foo, bar", ["foo", "bar"]),
+    ("foo, bar{1,3}", ["foo", "bar{1,3}"]),
+]
+
+
+@pytest.mark.parametrize("in_string,expected", CSV_REGEX_COMMA_CASES)
+def test_csv_regex_comma_in_quantifier(in_string, expected) -> None:
+    """Check that we correctly parse a comma-separated regex when there are one
+    or more commas within quantifier expressions.
+    """
+
+    def _template_run(in_string):
+        r = Run(
+            [str(EMPTY_MODULE), rf"--bad-names-rgx={in_string}"],
+            exit=False,
+        )
+        return r.linter.config.bad_names_rgxs
+
+    assert _template_run(in_string) == [re.compile(regex) for regex in expected]
+
+
+
 def test_short_verbose(capsys: CaptureFixture) -> None:
     """Check that we correctly handle the -v flag."""
     Run([str(EMPTY_MODULE), "-v"], exit=False)