bpo-30277: Replace _sre.getlower() with _sre.ascii_tolower() and _sre.unicode_tolower(). (#1468)
diff --git a/Lib/sre_compile.py b/Lib/sre_compile.py
index d7ee4e8..db8b8a2 100644
--- a/Lib/sre_compile.py
+++ b/Lib/sre_compile.py
@@ -69,13 +69,14 @@
REPEATING_CODES = _REPEATING_CODES
SUCCESS_CODES = _SUCCESS_CODES
ASSERT_CODES = _ASSERT_CODES
- if (flags & SRE_FLAG_IGNORECASE and
- not (flags & SRE_FLAG_LOCALE) and
- flags & SRE_FLAG_UNICODE and
- not (flags & SRE_FLAG_ASCII)):
- fixes = _ignorecase_fixes
- else:
- fixes = None
+ tolower = None
+ fixes = None
+ if flags & SRE_FLAG_IGNORECASE and not flags & SRE_FLAG_LOCALE:
+ if flags & SRE_FLAG_UNICODE and not flags & SRE_FLAG_ASCII:
+ tolower = _sre.unicode_tolower
+ fixes = _ignorecase_fixes
+ else:
+ tolower = _sre.ascii_tolower
for op, av in pattern:
if op in LITERAL_CODES:
if not flags & SRE_FLAG_IGNORECASE:
@@ -85,7 +86,7 @@
emit(OP_LOC_IGNORE[op])
emit(av)
else:
- lo = _sre.getlower(av, flags)
+ lo = tolower(av)
if fixes and lo in fixes:
emit(IN_IGNORE)
skip = _len(code); emit(0)
@@ -102,16 +103,12 @@
elif op is IN:
if not flags & SRE_FLAG_IGNORECASE:
emit(op)
- fixup = None
elif flags & SRE_FLAG_LOCALE:
emit(IN_LOC_IGNORE)
- fixup = None
else:
emit(IN_IGNORE)
- def fixup(literal, flags=flags):
- return _sre.getlower(literal, flags)
skip = _len(code); emit(0)
- _compile_charset(av, flags, code, fixup, fixes)
+ _compile_charset(av, flags, code, tolower, fixes)
code[skip] = _len(code) - skip
elif op is ANY:
if flags & SRE_FLAG_DOTALL:
diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py
index 7601dc8..b5b7cff 100644
--- a/Lib/test/test_re.py
+++ b/Lib/test/test_re.py
@@ -883,17 +883,23 @@
def test_category(self):
self.assertEqual(re.match(r"(\s)", " ").group(1), " ")
- def test_getlower(self):
+ @cpython_only
+ def test_case_helpers(self):
import _sre
- self.assertEqual(_sre.getlower(ord('A'), 0), ord('a'))
- self.assertEqual(_sre.getlower(ord('A'), re.LOCALE), ord('a'))
- self.assertEqual(_sre.getlower(ord('A'), re.UNICODE), ord('a'))
- self.assertEqual(_sre.getlower(ord('A'), re.ASCII), ord('a'))
+ 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)
- self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
- self.assertEqual(re.match(b"abc", b"ABC", re.I).group(0), b"ABC")
- self.assertEqual(re.match("abc", "ABC", re.I|re.A).group(0), "ABC")
- self.assertEqual(re.match(b"abc", b"ABC", re.I|re.L).group(0), b"ABC")
+ 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()))
+
+ self.assertEqual(_sre.ascii_tolower(0x0130), 0x0130)
+ self.assertEqual(_sre.unicode_tolower(0x0130), ord('i'))
def test_not_literal(self):
self.assertEqual(re.search(r"\s([^a])", " b").group(1), "b")