From 305ccbe27ea5ba82fd2d8c32ec739f980e524330 Mon Sep 17 00:00:00 2001 From: Serhiy Storchaka Date: Wed, 10 May 2017 06:05:20 +0300 Subject: [PATCH] bpo-30298: Weaken the condition of deprecation warnings for inline modifiers. (#1490) Now allowed several subsequential inline modifiers at the start of the pattern (e.g. '(?i)(?s)...'). In verbose mode whitespaces and comments now are allowed before and between inline modifiers (e.g. '(?x) (?i) (?s)...'). --- Lib/sre_parse.py | 12 ++++----- Lib/test/test_re.py | 59 ++++++++++++++++++++++++++++++++++----------- Misc/NEWS | 6 +++++ 3 files changed, 56 insertions(+), 21 deletions(-) diff --git a/Lib/sre_parse.py b/Lib/sre_parse.py index ab37fd3fe2f..d8d1bd552fb 100644 --- a/Lib/sre_parse.py +++ b/Lib/sre_parse.py @@ -412,7 +412,7 @@ def _parse_sub(source, state, verbose, nested=True): sourcematch = source.match start = source.tell() while True: - itemsappend(_parse(source, state, verbose)) + itemsappend(_parse(source, state, verbose, not nested and not items)) if not sourcematch("|"): break @@ -466,7 +466,7 @@ def _parse_sub_cond(source, state, condgroup, verbose): subpattern.append((GROUPREF_EXISTS, (condgroup, item_yes, item_no))) return subpattern -def _parse(source, state, verbose): +def _parse(source, state, verbose, first=False): # parse a simple pattern subpattern = SubPattern(state) @@ -730,10 +730,9 @@ def _parse(source, state, verbose): state.checklookbehindgroup(condgroup, source) elif char in FLAGS or char == "-": # flags - pos = source.pos flags = _parse_flags(source, state, char) if flags is None: # global flags - if pos != 3: # "(?x" + if not first or subpattern: import warnings warnings.warn( 'Flags not at the start of the expression %s%s' % ( @@ -742,6 +741,8 @@ def _parse(source, state, verbose): ), DeprecationWarning, stacklevel=7 ) + if (state.flags & SRE_FLAG_VERBOSE) and not verbose: + raise Verbose continue add_flags, del_flags = flags group = None @@ -795,9 +796,6 @@ def _parse_flags(source, state, char): msg = "unknown flag" if char.isalpha() else "missing -, : or )" raise source.error(msg, len(char)) if char == ")": - if ((add_flags & SRE_FLAG_VERBOSE) and - not (state.flags & SRE_FLAG_VERBOSE)): - raise Verbose state.flags |= add_flags return None if add_flags & GLOBAL_FLAGS: diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py index 3129f7e9888..4d71eea517e 100644 --- a/Lib/test/test_re.py +++ b/Lib/test/test_re.py @@ -1325,32 +1325,43 @@ class ReTests(unittest.TestCase): upper_char = '\u1ea0' # Latin Capital Letter A with Dot Below lower_char = '\u1ea1' # Latin Small Letter A with Dot Below - p = re.compile(upper_char, re.I | re.U) - q = p.match(lower_char) + p = re.compile('.' + upper_char, re.I | re.S) + q = p.match('\n' + lower_char) self.assertTrue(q) - p = re.compile(lower_char, re.I | re.U) - q = p.match(upper_char) + p = re.compile('.' + lower_char, re.I | re.S) + q = p.match('\n' + upper_char) self.assertTrue(q) - p = re.compile('(?i)' + upper_char, re.U) - q = p.match(lower_char) + p = re.compile('(?i).' + upper_char, re.S) + q = p.match('\n' + lower_char) self.assertTrue(q) - p = re.compile('(?i)' + lower_char, re.U) - q = p.match(upper_char) + p = re.compile('(?i).' + lower_char, re.S) + q = p.match('\n' + upper_char) self.assertTrue(q) - p = re.compile('(?iu)' + upper_char) - q = p.match(lower_char) + p = re.compile('(?is).' + upper_char) + q = p.match('\n' + lower_char) self.assertTrue(q) - p = re.compile('(?iu)' + lower_char) - q = p.match(upper_char) + p = re.compile('(?is).' + lower_char) + q = p.match('\n' + upper_char) self.assertTrue(q) - self.assertTrue(re.match('(?ixu) ' + upper_char, lower_char)) - self.assertTrue(re.match('(?ixu) ' + lower_char, upper_char)) + p = re.compile('(?s)(?i).' + upper_char) + q = p.match('\n' + lower_char) + self.assertTrue(q) + + p = re.compile('(?s)(?i).' + lower_char) + q = p.match('\n' + upper_char) + self.assertTrue(q) + + self.assertTrue(re.match('(?ix) ' + upper_char, lower_char)) + self.assertTrue(re.match('(?ix) ' + lower_char, upper_char)) + self.assertTrue(re.match(' (?i) ' + upper_char, lower_char, re.X)) + self.assertTrue(re.match('(?x) (?i) ' + upper_char, lower_char)) + self.assertTrue(re.match(' (?x) (?i) ' + upper_char, lower_char, re.X)) p = upper_char + '(?i)' with self.assertWarns(DeprecationWarning) as warns: @@ -1368,6 +1379,26 @@ class ReTests(unittest.TestCase): 'Flags not at the start of the expression %s (truncated)' % p[:20] ) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.match('(?s).(?i)' + upper_char, '\n' + lower_char)) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.match('(?i) ' + upper_char + ' (?x)', lower_char)) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.match(' (?x) (?i) ' + upper_char, lower_char)) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.match('^(?i)' + upper_char, lower_char)) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.match('$|(?i)' + upper_char, lower_char)) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.match('(?:(?i)' + upper_char + ')', lower_char)) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.fullmatch('(^)?(?(1)(?i)' + upper_char + ')', + lower_char)) + with self.assertWarns(DeprecationWarning): + self.assertTrue(re.fullmatch('($)?(?(1)|(?i)' + upper_char + ')', + lower_char)) + + def test_dollar_matches_twice(self): "$ matches the end of string, and just before the terminating \n" pattern = re.compile('$') diff --git a/Misc/NEWS b/Misc/NEWS index 7a79521efd7..997a03411de 100644 --- a/Misc/NEWS +++ b/Misc/NEWS @@ -320,6 +320,12 @@ Extension Modules Library ------- +- bpo-30298: Weaken the condition of deprecation warnings for inline modifiers. + Now allowed several subsequential inline modifiers at the start of the + pattern (e.g. ``'(?i)(?s)...'``). In verbose mode whitespaces and comments + now are allowed before and between inline modifiers (e.g. + ``'(?x) (?i) (?s)...'``). + - bpo-30285: Optimized case-insensitive matching and searching of regular expressions.