https://github.com/python/cpython/commit/cd6abe27a2582786da7b4b8bb008910563f31735 commit: cd6abe27a2582786da7b4b8bb008910563f31735 branch: main author: Andrew Shteren <[email protected]> committer: jaraco <[email protected]> date: 2025-02-24T00:20:37Z summary:
gh-128641: Fix ConfigParser.read Perfomance Regression (#129596) --------- Co-authored-by: Jason R. Coombs <[email protected]> Co-authored-by: Bénédikt Tran <[email protected]> files: A Misc/NEWS.d/next/Library/2025-02-02-23-47-35.gh-issue-128641.GFs673.rst M Lib/configparser.py diff --git a/Lib/configparser.py b/Lib/configparser.py index 462af2f4abf867..70cc651edabd86 100644 --- a/Lib/configparser.py +++ b/Lib/configparser.py @@ -154,7 +154,6 @@ import os import re import sys -import types __all__ = ("NoSectionError", "DuplicateOptionError", "DuplicateSectionError", "NoOptionError", "InterpolationError", "InterpolationDepthError", @@ -570,35 +569,36 @@ def __init__(self): class _Line(str): + __slots__ = 'clean', 'has_comments' def __new__(cls, val, *args, **kwargs): return super().__new__(cls, val) - def __init__(self, val, prefixes): - self.prefixes = prefixes + def __init__(self, val, comments): + trimmed = val.strip() + self.clean = comments.strip(trimmed) + self.has_comments = trimmed != self.clean - @functools.cached_property - def clean(self): - return self._strip_full() and self._strip_inline() - @property - def has_comments(self): - return self.strip() != self.clean - - def _strip_inline(self): - """ - Search for the earliest prefix at the beginning of the line or following a space. - """ - matcher = re.compile( - '|'.join(fr'(^|\s)({re.escape(prefix)})' for prefix in self.prefixes.inline) - # match nothing if no prefixes - or '(?!)' +class _CommentSpec: + def __init__(self, full_prefixes, inline_prefixes): + full_patterns = ( + # prefix at the beginning of a line + fr'^({re.escape(prefix)}).*' + for prefix in full_prefixes ) - match = matcher.search(self) - return self[:match.start() if match else None].strip() + inline_patterns = ( + # prefix at the beginning of the line or following a space + fr'(^|\s)({re.escape(prefix)}.*)' + for prefix in inline_prefixes + ) + self.pattern = re.compile('|'.join(itertools.chain(full_patterns, inline_patterns))) + + def strip(self, text): + return self.pattern.sub('', text).rstrip() - def _strip_full(self): - return '' if any(map(self.strip().startswith, self.prefixes.full)) else True + def wrap(self, text): + return _Line(text, self) class RawConfigParser(MutableMapping): @@ -667,10 +667,7 @@ def __init__(self, defaults=None, dict_type=_default_dict, else: self._optcre = re.compile(self._OPT_TMPL.format(delim=d), re.VERBOSE) - self._prefixes = types.SimpleNamespace( - full=tuple(comment_prefixes or ()), - inline=tuple(inline_comment_prefixes or ()), - ) + self._comments = _CommentSpec(comment_prefixes or (), inline_comment_prefixes or ()) self._strict = strict self._allow_no_value = allow_no_value self._empty_lines_in_values = empty_lines_in_values @@ -1066,7 +1063,6 @@ def _read(self, fp, fpname): in an otherwise empty line or may be entered in lines holding values or section names. Please note that comments get stripped off when reading configuration files. """ - try: ParsingError._raise_all(self._read_inner(fp, fpname)) finally: @@ -1075,8 +1071,7 @@ def _read(self, fp, fpname): def _read_inner(self, fp, fpname): st = _ReadState() - Line = functools.partial(_Line, prefixes=self._prefixes) - for st.lineno, line in enumerate(map(Line, fp), start=1): + for st.lineno, line in enumerate(map(self._comments.wrap, fp), start=1): if not line.clean: if self._empty_lines_in_values: # add empty line to the value, but only if there was no diff --git a/Misc/NEWS.d/next/Library/2025-02-02-23-47-35.gh-issue-128641.GFs673.rst b/Misc/NEWS.d/next/Library/2025-02-02-23-47-35.gh-issue-128641.GFs673.rst new file mode 100644 index 00000000000000..bfc0f8fe6d04fb --- /dev/null +++ b/Misc/NEWS.d/next/Library/2025-02-02-23-47-35.gh-issue-128641.GFs673.rst @@ -0,0 +1 @@ +Restore :meth:`configparser.ConfigParser.read` performance. _______________________________________________ Python-checkins mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3/lists/python-checkins.python.org/ Member address: [email protected]
