[3.11] Move implementation specific RE tests to separate class (GH-106563) (GH-106565)

(cherry picked from commit 8cb6f9761e)

Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
This commit is contained in:
Miss Islington (bot) 2023-07-09 05:23:27 -07:00 committed by GitHub
parent 2d037fb406
commit 769b7d2d0b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23

View file

@ -1077,33 +1077,6 @@ class ReTests(unittest.TestCase):
def test_category(self):
self.assertEqual(re.match(r"(\s)", " ").group(1), " ")
@cpython_only
def test_case_helpers(self):
import _sre
for i in range(128):
c = chr(i)
lo = ord(c.lower())
self.assertEqual(_sre.ascii_tolower(i), lo)
self.assertEqual(_sre.unicode_tolower(i), lo)
iscased = c in string.ascii_letters
self.assertEqual(_sre.ascii_iscased(i), iscased)
self.assertEqual(_sre.unicode_iscased(i), iscased)
for i in list(range(128, 0x1000)) + [0x10400, 0x10428]:
c = chr(i)
self.assertEqual(_sre.ascii_tolower(i), i)
if i != 0x0130:
self.assertEqual(_sre.unicode_tolower(i), ord(c.lower()))
iscased = c != c.lower() or c != c.upper()
self.assertFalse(_sre.ascii_iscased(i))
self.assertEqual(_sre.unicode_iscased(i),
c != c.lower() or c != c.upper())
self.assertEqual(_sre.ascii_tolower(0x0130), 0x0130)
self.assertEqual(_sre.unicode_tolower(0x0130), ord('i'))
self.assertFalse(_sre.ascii_iscased(0x0130))
self.assertTrue(_sre.unicode_iscased(0x0130))
def test_not_literal(self):
self.assertEqual(re.search(r"\s([^a])", " b").group(1), "b")
self.assertEqual(re.search(r"\s([^a]*)", " bb").group(1), "bb")
@ -1800,20 +1773,6 @@ class ReTests(unittest.TestCase):
pat = re.compile(b'..')
self.assertEqual(pat.sub(lambda m: b'bytes', b'a5'), b'bytes')
def test_dealloc(self):
# issue 3299: check for segfault in debug build
import _sre
# the overflow limit is different on wide and narrow builds and it
# depends on the definition of SRE_CODE (see sre.h).
# 2**128 should be big enough to overflow on both. For smaller values
# a RuntimeError is raised instead of OverflowError.
long_overflow = 2**128
self.assertRaises(TypeError, re.finditer, "a", {})
with self.assertRaises(OverflowError):
_sre.compile("abc", 0, [long_overflow], 0, {}, ())
with self.assertRaises(TypeError):
_sre.compile({}, 0, [], 0, [], [])
def test_search_dot_unicode(self):
self.assertTrue(re.search("123.*-", '123abc-'))
self.assertTrue(re.search("123.*-", '123\xe9-'))
@ -1871,21 +1830,6 @@ class ReTests(unittest.TestCase):
self.assertRaises(OverflowError, re.compile, r".{%d,}?" % 2**128)
self.assertRaises(OverflowError, re.compile, r".{%d,%d}" % (2**129, 2**128))
@cpython_only
def test_repeat_minmax_overflow_maxrepeat(self):
try:
from _sre import MAXREPEAT
except ImportError:
self.skipTest('requires _sre.MAXREPEAT constant')
string = "x" * 100000
self.assertIsNone(re.match(r".{%d}" % (MAXREPEAT - 1), string))
self.assertEqual(re.match(r".{,%d}" % (MAXREPEAT - 1), string).span(),
(0, 100000))
self.assertIsNone(re.match(r".{%d,}?" % (MAXREPEAT - 1), string))
self.assertRaises(OverflowError, re.compile, r".{%d}" % MAXREPEAT)
self.assertRaises(OverflowError, re.compile, r".{,%d}" % MAXREPEAT)
self.assertRaises(OverflowError, re.compile, r".{%d,}?" % MAXREPEAT)
def test_backref_group_name_in_exception(self):
# Issue 17341: Poor error message when compiling invalid regex
self.checkPatternError('(?P=<foo>)',
@ -2720,6 +2664,64 @@ class ImplementationTest(unittest.TestCase):
self.assertTrue(hasattr(mod, attr))
del sys.modules[name]
@cpython_only
def test_case_helpers(self):
import _sre
for i in range(128):
c = chr(i)
lo = ord(c.lower())
self.assertEqual(_sre.ascii_tolower(i), lo)
self.assertEqual(_sre.unicode_tolower(i), lo)
iscased = c in string.ascii_letters
self.assertEqual(_sre.ascii_iscased(i), iscased)
self.assertEqual(_sre.unicode_iscased(i), iscased)
for i in list(range(128, 0x1000)) + [0x10400, 0x10428]:
c = chr(i)
self.assertEqual(_sre.ascii_tolower(i), i)
if i != 0x0130:
self.assertEqual(_sre.unicode_tolower(i), ord(c.lower()))
iscased = c != c.lower() or c != c.upper()
self.assertFalse(_sre.ascii_iscased(i))
self.assertEqual(_sre.unicode_iscased(i),
c != c.lower() or c != c.upper())
self.assertEqual(_sre.ascii_tolower(0x0130), 0x0130)
self.assertEqual(_sre.unicode_tolower(0x0130), ord('i'))
self.assertFalse(_sre.ascii_iscased(0x0130))
self.assertTrue(_sre.unicode_iscased(0x0130))
@cpython_only
def test_dealloc(self):
# issue 3299: check for segfault in debug build
import _sre
# the overflow limit is different on wide and narrow builds and it
# depends on the definition of SRE_CODE (see sre.h).
# 2**128 should be big enough to overflow on both. For smaller values
# a RuntimeError is raised instead of OverflowError.
long_overflow = 2**128
self.assertRaises(TypeError, re.finditer, "a", {})
with self.assertRaises(OverflowError):
_sre.compile("abc", 0, [long_overflow], 0, {}, ())
with self.assertRaises(TypeError):
_sre.compile({}, 0, [], 0, [], [])
@cpython_only
def test_repeat_minmax_overflow_maxrepeat(self):
try:
from _sre import MAXREPEAT
except ImportError:
self.skipTest('requires _sre.MAXREPEAT constant')
string = "x" * 100000
self.assertIsNone(re.match(r".{%d}" % (MAXREPEAT - 1), string))
self.assertEqual(re.match(r".{,%d}" % (MAXREPEAT - 1), string).span(),
(0, 100000))
self.assertIsNone(re.match(r".{%d,}?" % (MAXREPEAT - 1), string))
self.assertRaises(OverflowError, re.compile, r".{%d}" % MAXREPEAT)
self.assertRaises(OverflowError, re.compile, r".{,%d}" % MAXREPEAT)
self.assertRaises(OverflowError, re.compile, r".{%d,}?" % MAXREPEAT)
class ExternalTests(unittest.TestCase):
def test_re_benchmarks(self):