traverseda
/
cpython
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Issue #12728: Different Unicode characters having the same uppercase but 

different lowercase are now matched in case-insensitive regular expressions.
This commit is contained in:
Serhiy Storchaka 2014-11-10 12:37:16 +02:00
parent d915b0847d
commit 0c938f6d24
3 changed files with 135 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

92
Lib/sre_compile.py
View File

@ -27,6 +27,46 @@ _REPEATING_CODES = set([REPEAT, MIN_REPEAT, MAX_REPEAT])
_SUCCESS_CODES = set([SUCCESS, FAILURE])
_ASSERT_CODES = set([ASSERT, ASSERT_NOT])
# Sets of lowercase characters which have the same uppercase.
_equivalences = (
# LATIN SMALL LETTER I, LATIN SMALL LETTER DOTLESS I
(0x69, 0x131), # iı
# LATIN SMALL LETTER S, LATIN SMALL LETTER LONG S
(0x73, 0x17f), # sſ
# MICRO SIGN, GREEK SMALL LETTER MU
(0xb5, 0x3bc), # µμ
# COMBINING GREEK YPOGEGRAMMENI, GREEK SMALL LETTER IOTA, GREEK PROSGEGRAMMENI
(0x345, 0x3b9, 0x1fbe), # \u0345ιι
# GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS, GREEK SMALL LETTER IOTA WITH DIALYTIKA AND OXIA
(0x390, 0x1fd3), # ΐΐ
# GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS, GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND OXIA
(0x3b0, 0x1fe3), # ΰΰ
# GREEK SMALL LETTER BETA, GREEK BETA SYMBOL
(0x3b2, 0x3d0), # βϐ
# GREEK SMALL LETTER EPSILON, GREEK LUNATE EPSILON SYMBOL
(0x3b5, 0x3f5), # εϵ
# GREEK SMALL LETTER THETA, GREEK THETA SYMBOL
(0x3b8, 0x3d1), # θϑ
# GREEK SMALL LETTER KAPPA, GREEK KAPPA SYMBOL
(0x3ba, 0x3f0), # κϰ
# GREEK SMALL LETTER PI, GREEK PI SYMBOL
(0x3c0, 0x3d6), # πϖ
# GREEK SMALL LETTER RHO, GREEK RHO SYMBOL
(0x3c1, 0x3f1), # ρϱ
# GREEK SMALL LETTER FINAL SIGMA, GREEK SMALL LETTER SIGMA
(0x3c2, 0x3c3), # ςσ
# GREEK SMALL LETTER PHI, GREEK PHI SYMBOL
(0x3c6, 0x3d5), # φϕ
# LATIN SMALL LETTER S WITH DOT ABOVE, LATIN SMALL LETTER LONG S WITH DOT ABOVE
(0x1e61, 0x1e9b), # ṡẛ
# LATIN SMALL LIGATURE LONG S T, LATIN SMALL LIGATURE ST
(0xfb05, 0xfb06), # ſtst
)
# Maps the lowercase code to lowercase codes which have the same uppercase.
_ignorecase_fixes = {i: tuple(j for j in t if i != j)
for t in _equivalences for i in t}
def _compile(code, pattern, flags):
# internal: compile a (sub)pattern
emit = code.append
@ -35,11 +75,29 @@ def _compile(code, pattern, flags):
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):
fixes = _ignorecase_fixes
else:
fixes = None
for op, av in pattern:
if op in LITERAL_CODES:
if flags & SRE_FLAG_IGNORECASE:
emit(OPCODES[OP_IGNORE[op]])
emit(_sre.getlower(av, flags))
lo = _sre.getlower(av, flags)
if fixes and lo in fixes:
emit(OPCODES[IN_IGNORE])
skip = _len(code); emit(0)
if op is NOT_LITERAL:
emit(OPCODES[NEGATE])
for k in (lo,) + fixes[lo]:
emit(OPCODES[LITERAL])
emit(k)
emit(OPCODES[FAILURE])
code[skip] = _len(code) - skip
else:
emit(OPCODES[OP_IGNORE[op]])
emit(lo)
else:
emit(OPCODES[op])
emit(av)
@ -52,7 +110,7 @@ def _compile(code, pattern, flags):
emit(OPCODES[op])
fixup = None
skip = _len(code); emit(0)
_compile_charset(av, flags, code, fixup)
_compile_charset(av, flags, code, fixup, fixes)
code[skip] = _len(code) - skip
elif op is ANY:
if flags & SRE_FLAG_DOTALL:
@ -166,10 +224,11 @@ def _compile(code, pattern, flags):
else:
raise ValueError("unsupported operand type", op)
def _compile_charset(charset, flags, code, fixup=None):
def _compile_charset(charset, flags, code, fixup=None, fixes=None):
# compile charset subprogram
emit = code.append
for op, av in _optimize_charset(charset, fixup, flags & SRE_FLAG_UNICODE):
for op, av in _optimize_charset(charset, fixup, fixes,
flags & SRE_FLAG_UNICODE):
emit(OPCODES[op])
if op is NEGATE:
pass
@ -193,7 +252,7 @@ def _compile_charset(charset, flags, code, fixup=None):
raise error("internal: unsupported set operator")
emit(OPCODES[FAILURE])
def _optimize_charset(charset, fixup, isunicode):
def _optimize_charset(charset, fixup, fixes, isunicode):
# internal: optimize character set
out = []
tail = []
@ -202,16 +261,27 @@ def _optimize_charset(charset, fixup, isunicode):
while True:
try:
if op is LITERAL:
i = av
if fixup:
i = fixup(i)
charmap[i] = 1
i = fixup(av)
charmap[i] = 1
if fixes and i in fixes:
for k in fixes[i]:
charmap[k] = 1
else:
charmap[av] = 1
elif op is RANGE:
r = range(av[0], av[1]+1)
if fixup:
r = map(fixup, r)
for i in r:
charmap[i] = 1
if fixup and fixes:
for i in r:
charmap[i] = 1
if i in fixes:
for k in fixes[i]:
charmap[k] = 1
else:
for i in r:
charmap[i] = 1
elif op is NEGATE:
out.append((op, av))
else:

51
Lib/test/test_re.py
View File

@ -619,6 +619,43 @@ class ReTests(unittest.TestCase):
self.assertEqual(re.match(r"((a)\s(abc|a))", "a a", re.I).group(1), "a a")
self.assertEqual(re.match(r"((a)\s(abc|a)*)", "a aa", re.I).group(1), "a aa")
assert '\u212a'.lower() == 'k' # 'K'
self.assertTrue(re.match(r'K', '\u212a', re.I))
self.assertTrue(re.match(r'k', '\u212a', re.I))
self.assertTrue(re.match(r'\u212a', 'K', re.I))
self.assertTrue(re.match(r'\u212a', 'k', re.I))
assert '\u017f'.upper() == 'S' # 'ſ'
self.assertTrue(re.match(r'S', '\u017f', re.I))
self.assertTrue(re.match(r's', '\u017f', re.I))
self.assertTrue(re.match(r'\u017f', 'S', re.I))
self.assertTrue(re.match(r'\u017f', 's', re.I))
assert '\ufb05'.upper() == '\ufb06'.upper() == 'ST' # 'ſt', 'st'
self.assertTrue(re.match(r'\ufb05', '\ufb06', re.I))
self.assertTrue(re.match(r'\ufb06', '\ufb05', re.I))
def test_ignore_case_set(self):
self.assertTrue(re.match(r'[19A]', 'A', re.I))
self.assertTrue(re.match(r'[19a]', 'a', re.I))
self.assertTrue(re.match(r'[19a]', 'A', re.I))
self.assertTrue(re.match(r'[19A]', 'a', re.I))
self.assertTrue(re.match(br'[19A]', b'A', re.I))
self.assertTrue(re.match(br'[19a]', b'a', re.I))
self.assertTrue(re.match(br'[19a]', b'A', re.I))
self.assertTrue(re.match(br'[19A]', b'a', re.I))
assert '\u212a'.lower() == 'k' # 'K'
self.assertTrue(re.match(r'[19K]', '\u212a', re.I))
self.assertTrue(re.match(r'[19k]', '\u212a', re.I))
self.assertTrue(re.match(r'[19\u212a]', 'K', re.I))
self.assertTrue(re.match(r'[19\u212a]', 'k', re.I))
assert '\u017f'.upper() == 'S' # 'ſ'
self.assertTrue(re.match(r'[19S]', '\u017f', re.I))
self.assertTrue(re.match(r'[19s]', '\u017f', re.I))
self.assertTrue(re.match(r'[19\u017f]', 'S', re.I))
self.assertTrue(re.match(r'[19\u017f]', 's', re.I))
assert '\ufb05'.upper() == '\ufb06'.upper() == 'ST' # 'ſt', 'st'
self.assertTrue(re.match(r'[19\ufb05]', '\ufb06', re.I))
self.assertTrue(re.match(r'[19\ufb06]', '\ufb05', re.I))
def test_ignore_case_range(self):
# Issues #3511, #17381.
self.assertTrue(re.match(r'[9-a]', '_', re.I))
@ -638,6 +675,20 @@ class ReTests(unittest.TestCase):
self.assertTrue(re.match(r'[\U00010400-\U00010427]', '\U00010428', re.I))
self.assertTrue(re.match(r'[\U00010400-\U00010427]', '\U00010400', re.I))
assert '\u212a'.lower() == 'k' # 'K'
self.assertTrue(re.match(r'[J-M]', '\u212a', re.I))
self.assertTrue(re.match(r'[j-m]', '\u212a', re.I))
self.assertTrue(re.match(r'[\u2129-\u212b]', 'K', re.I))
self.assertTrue(re.match(r'[\u2129-\u212b]', 'k', re.I))
assert '\u017f'.upper() == 'S' # 'ſ'
self.assertTrue(re.match(r'[R-T]', '\u017f', re.I))
self.assertTrue(re.match(r'[r-t]', '\u017f', re.I))
self.assertTrue(re.match(r'[\u017e-\u0180]', 'S', re.I))
self.assertTrue(re.match(r'[\u017e-\u0180]', 's', re.I))
assert '\ufb05'.upper() == '\ufb06'.upper() == 'ST' # 'ſt', 'st'
self.assertTrue(re.match(r'[\ufb04-\ufb05]', '\ufb06', re.I))
self.assertTrue(re.match(r'[\ufb06-\ufb07]', '\ufb05', re.I))
def test_category(self):
self.assertEqual(re.match(r"(\s)", " ").group(1), " ")

3
Misc/NEWS
View File

@ -36,6 +36,9 @@ Core and Builtins
Library
-------
- Issue #12728: Different Unicode characters having the same uppercase but
different lowercase are now matched in case-insensitive regular expressions.
- Issue #22821: Fixed fcntl() with integer argument on 64-bit big-endian
platforms.