summaryrefslogtreecommitdiffstats
path: root/Lib/sre_compile.py
diff options
context:
space:
mode:
authorSerhiy Storchaka <storchaka@gmail.com>2014-11-10 10:37:16 (GMT)
committerSerhiy Storchaka <storchaka@gmail.com>2014-11-10 10:37:16 (GMT)
commit0c938f6d248e7dbf50a1f00b0c6bca3845859fc0 (patch)
treea8ff55cea441decd74a0c514d0c1152789eed682 /Lib/sre_compile.py
parentd915b0847dd5e43b33cf0683bfc473e904572a56 (diff)
downloadcpython-0c938f6d248e7dbf50a1f00b0c6bca3845859fc0.zip
cpython-0c938f6d248e7dbf50a1f00b0c6bca3845859fc0.tar.gz
cpython-0c938f6d248e7dbf50a1f00b0c6bca3845859fc0.tar.bz2
Issue #12728: Different Unicode characters having the same uppercase but
different lowercase are now matched in case-insensitive regular expressions.
Diffstat (limited to 'Lib/sre_compile.py')
-rw-r--r--Lib/sre_compile.py92
1 files changed, 81 insertions, 11 deletions
diff --git a/Lib/sre_compile.py b/Lib/sre_compile.py
index f24f681..53baa0d 100644
--- a/Lib/sre_compile.py
+++ b/Lib/sre_compile.py
@@ -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: