summaryrefslogtreecommitdiffstats
path: root/Lib
diff options
context:
space:
mode:
authorSKO <41810398+uyw4687@users.noreply.github.com>2023-08-16 07:43:45 (GMT)
committerGitHub <noreply@github.com>2023-08-16 07:43:45 (GMT)
commitabd9cc52d94b8e2835322b62c29f09bb0e6fcfe9 (patch)
tree252a5f9d86fa4746cf7615ab7e6f16ceb12dede5 /Lib
parenta86df298df5b02e2d69ea6879e9ed10a7adb85d0 (diff)
downloadcpython-abd9cc52d94b8e2835322b62c29f09bb0e6fcfe9.zip
cpython-abd9cc52d94b8e2835322b62c29f09bb0e6fcfe9.tar.gz
cpython-abd9cc52d94b8e2835322b62c29f09bb0e6fcfe9.tar.bz2
gh-100061: Proper fix of the bug in the matching of possessive quantifiers (GH-102612)
Restore the global Input Stream pointer after trying to match a sub-pattern. Co-authored-by: Ma Lin <animalize@users.noreply.github.com>
Diffstat (limited to 'Lib')
-rw-r--r--Lib/re/_compiler.py7
-rw-r--r--Lib/test/test_re.py14
2 files changed, 11 insertions, 10 deletions
diff --git a/Lib/re/_compiler.py b/Lib/re/_compiler.py
index f5fd160..d0a4c55 100644
--- a/Lib/re/_compiler.py
+++ b/Lib/re/_compiler.py
@@ -100,13 +100,6 @@ def _compile(code, pattern, flags):
emit(ANY_ALL)
else:
emit(ANY)
- elif op is POSSESSIVE_REPEAT:
- # gh-106052: Possessive quantifiers do not work when the
- # subpattern contains backtracking, i.e. "(?:ab?c)*+".
- # Implement it as equivalent greedy qualifier in atomic group.
- p = [(MAX_REPEAT, av)]
- p = [(ATOMIC_GROUP, p)]
- _compile(code, p, flags)
elif op in REPEATING_CODES:
if _simple(av[2]):
emit(REPEATING_CODES[op][2])
diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py
index bf3698a..042f97f 100644
--- a/Lib/test/test_re.py
+++ b/Lib/test/test_re.py
@@ -2342,7 +2342,17 @@ class ReTests(unittest.TestCase):
self.assertTrue(re.fullmatch(r'(?s:(?>.*?\.).*)\Z', "a.txt")) # reproducer
self.assertTrue(re.fullmatch(r'(?s:(?=(?P<g0>.*?\.))(?P=g0).*)\Z', "a.txt"))
- def test_bug_gh106052(self):
+ def test_bug_gh100061(self):
+ # gh-100061
+ self.assertEqual(re.match('(?>(?:.(?!D))+)', 'ABCDE').span(), (0, 2))
+ self.assertEqual(re.match('(?:.(?!D))++', 'ABCDE').span(), (0, 2))
+ self.assertEqual(re.match('(?>(?:.(?!D))*)', 'ABCDE').span(), (0, 2))
+ self.assertEqual(re.match('(?:.(?!D))*+', 'ABCDE').span(), (0, 2))
+ self.assertEqual(re.match('(?>(?:.(?!D))?)', 'CDE').span(), (0, 0))
+ self.assertEqual(re.match('(?:.(?!D))?+', 'CDE').span(), (0, 0))
+ self.assertEqual(re.match('(?>(?:.(?!D)){1,3})', 'ABCDE').span(), (0, 2))
+ self.assertEqual(re.match('(?:.(?!D)){1,3}+', 'ABCDE').span(), (0, 2))
+ # gh-106052
self.assertEqual(re.match("(?>(?:ab?c)+)", "aca").span(), (0, 2))
self.assertEqual(re.match("(?:ab?c)++", "aca").span(), (0, 2))
self.assertEqual(re.match("(?>(?:ab?c)*)", "aca").span(), (0, 2))
@@ -2451,7 +2461,6 @@ ATOMIC_GROUP
17: SUCCESS
''')
- @unittest.expectedFailure # gh-106052
def test_possesive_repeat_one(self):
self.assertEqual(get_debug_out(r'a?+'), '''\
POSSESSIVE_REPEAT 0 1
@@ -2464,7 +2473,6 @@ POSSESSIVE_REPEAT 0 1
12: SUCCESS
''')
- @unittest.expectedFailure # gh-106052
def test_possesive_repeat(self):
self.assertEqual(get_debug_out(r'(?:ab)?+'), '''\
POSSESSIVE_REPEAT 0 1