Skip to content

Commit 819830f

Browse files
serhiy-storchakaserhiy-storchaka
authored
gh-126505: Fix bugs in compiling case-insensitive character classes (GH-126557)
* upper-case non-BMP character was ignored * the ASCII flag was ignored when matching a character range whose upper bound is beyond the BMP region
1 parent 9fc2808 commit 819830f

File tree

3 files changed

+73
-9
lines changed

3 files changed

+73
-9
lines changed

‎Lib/re/_compiler.py‎

Lines changed: 14 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -255,19 +255,19 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None):
255255
while True:
256256
try:
257257
if op is LITERAL:
258-
if fixup:
259-
lo = fixup(av)
260-
charmap[lo] = 1
261-
if fixes and lo in fixes:
262-
for k in fixes[lo]:
258+
if fixup: # IGNORECASE and not LOCALE
259+
av = fixup(av)
260+
charmap[av] = 1
261+
if fixes and av in fixes:
262+
for k in fixes[av]:
263263
charmap[k] = 1
264264
if not hascased and iscased(av):
265265
hascased = True
266266
else:
267267
charmap[av] = 1
268268
elif op is RANGE:
269269
r = range(av[0], av[1]+1)
270-
if fixup:
270+
if fixup: # IGNORECASE and not LOCALE
271271
if fixes:
272272
for i in map(fixup, r):
273273
charmap[i] = 1
@@ -298,8 +298,7 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None):
298298
# Character set contains non-BMP character codes.
299299
# For range, all BMP characters in the range are already
300300
# proceeded.
301-
if fixup:
302-
hascased = True
301+
if fixup: # IGNORECASE and not LOCALE
303302
# For now, IN_UNI_IGNORE+LITERAL and
304303
# IN_UNI_IGNORE+RANGE_UNI_IGNORE work for all non-BMP
305304
# characters, because two characters (at least one of
@@ -310,7 +309,13 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None):
310309
# Also, both c.lower() and c.lower().upper() are single
311310
# characters for every non-BMP character.
312311
if op is RANGE:
313-
op = RANGE_UNI_IGNORE
312+
if fixes: # not ASCII
313+
op = RANGE_UNI_IGNORE
314+
hascased = True
315+
else:
316+
assert op is LITERAL
317+
if not hascased and iscased(av):
318+
hascased = True
314319
tail.append((op, av))
315320
break
316321

‎Lib/test/test_re.py‎

Lines changed: 55 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1136,6 +1136,39 @@ def test_ignore_case_set(self):
11361136
self.assertTrue(re.match(br'[19a]', b'a', re.I))
11371137
self.assertTrue(re.match(br'[19a]', b'A', re.I))
11381138
self.assertTrue(re.match(br'[19A]', b'a', re.I))
1139+
self.assertTrue(re.match(r'[19\xc7]', '\xc7', re.I))
1140+
self.assertTrue(re.match(r'[19\xc7]', '\xe7', re.I))
1141+
self.assertTrue(re.match(r'[19\xe7]', '\xc7', re.I))
1142+
self.assertTrue(re.match(r'[19\xe7]', '\xe7', re.I))
1143+
self.assertTrue(re.match(r'[19\u0400]', '\u0400', re.I))
1144+
self.assertTrue(re.match(r'[19\u0400]', '\u0450', re.I))
1145+
self.assertTrue(re.match(r'[19\u0450]', '\u0400', re.I))
1146+
self.assertTrue(re.match(r'[19\u0450]', '\u0450', re.I))
1147+
self.assertTrue(re.match(r'[19\U00010400]', '\U00010400', re.I))
1148+
self.assertTrue(re.match(r'[19\U00010400]', '\U00010428', re.I))
1149+
self.assertTrue(re.match(r'[19\U00010428]', '\U00010400', re.I))
1150+
self.assertTrue(re.match(r'[19\U00010428]', '\U00010428', re.I))
1151+
1152+
self.assertTrue(re.match(br'[19A]', b'A', re.I))
1153+
self.assertTrue(re.match(br'[19a]', b'a', re.I))
1154+
self.assertTrue(re.match(br'[19a]', b'A', re.I))
1155+
self.assertTrue(re.match(br'[19A]', b'a', re.I))
1156+
self.assertTrue(re.match(r'[19A]', 'A', re.I|re.A))
1157+
self.assertTrue(re.match(r'[19a]', 'a', re.I|re.A))
1158+
self.assertTrue(re.match(r'[19a]', 'A', re.I|re.A))
1159+
self.assertTrue(re.match(r'[19A]', 'a', re.I|re.A))
1160+
self.assertTrue(re.match(r'[19\xc7]', '\xc7', re.I|re.A))
1161+
self.assertIsNone(re.match(r'[19\xc7]', '\xe7', re.I|re.A))
1162+
self.assertIsNone(re.match(r'[19\xe7]', '\xc7', re.I|re.A))
1163+
self.assertTrue(re.match(r'[19\xe7]', '\xe7', re.I|re.A))
1164+
self.assertTrue(re.match(r'[19\u0400]', '\u0400', re.I|re.A))
1165+
self.assertIsNone(re.match(r'[19\u0400]', '\u0450', re.I|re.A))
1166+
self.assertIsNone(re.match(r'[19\u0450]', '\u0400', re.I|re.A))
1167+
self.assertTrue(re.match(r'[19\u0450]', '\u0450', re.I|re.A))
1168+
self.assertTrue(re.match(r'[19\U00010400]', '\U00010400', re.I|re.A))
1169+
self.assertIsNone(re.match(r'[19\U00010400]', '\U00010428', re.I|re.A))
1170+
self.assertIsNone(re.match(r'[19\U00010428]', '\U00010400', re.I|re.A))
1171+
self.assertTrue(re.match(r'[19\U00010428]', '\U00010428', re.I|re.A))
11391172

11401173
# Two different characters have the same lowercase.
11411174
assert 'K'.lower() == '\u212a'.lower() == 'k' # 'K'
@@ -1172,8 +1205,10 @@ def test_ignore_case_range(self):
11721205
self.assertTrue(re.match(br'[9-a]', b'_', re.I))
11731206
self.assertIsNone(re.match(br'[9-A]', b'_', re.I))
11741207
self.assertTrue(re.match(r'[\xc0-\xde]', '\xd7', re.I))
1208+
self.assertTrue(re.match(r'[\xc0-\xde]', '\xe7', re.I))
11751209
self.assertIsNone(re.match(r'[\xc0-\xde]', '\xf7', re.I))
11761210
self.assertTrue(re.match(r'[\xe0-\xfe]', '\xf7', re.I))
1211+
self.assertTrue(re.match(r'[\xe0-\xfe]', '\xc7', re.I))
11771212
self.assertIsNone(re.match(r'[\xe0-\xfe]', '\xd7', re.I))
11781213
self.assertTrue(re.match(r'[\u0430-\u045f]', '\u0450', re.I))
11791214
self.assertTrue(re.match(r'[\u0430-\u045f]', '\u0400', re.I))
@@ -1184,6 +1219,26 @@ def test_ignore_case_range(self):
11841219
self.assertTrue(re.match(r'[\U00010400-\U00010427]', '\U00010428', re.I))
11851220
self.assertTrue(re.match(r'[\U00010400-\U00010427]', '\U00010400', re.I))
11861221

1222+
self.assertTrue(re.match(r'[\xc0-\xde]', '\xd7', re.I|re.A))
1223+
self.assertIsNone(re.match(r'[\xc0-\xde]', '\xe7', re.I|re.A))
1224+
self.assertTrue(re.match(r'[\xe0-\xfe]', '\xf7', re.I|re.A))
1225+
self.assertIsNone(re.match(r'[\xe0-\xfe]', '\xc7', re.I|re.A))
1226+
self.assertTrue(re.match(r'[\u0430-\u045f]', '\u0450', re.I|re.A))
1227+
self.assertIsNone(re.match(r'[\u0430-\u045f]', '\u0400', re.I|re.A))
1228+
self.assertIsNone(re.match(r'[\u0400-\u042f]', '\u0450', re.I|re.A))
1229+
self.assertTrue(re.match(r'[\u0400-\u042f]', '\u0400', re.I|re.A))
1230+
self.assertTrue(re.match(r'[\U00010428-\U0001044f]', '\U00010428', re.I|re.A))
1231+
self.assertIsNone(re.match(r'[\U00010428-\U0001044f]', '\U00010400', re.I|re.A))
1232+
self.assertIsNone(re.match(r'[\U00010400-\U00010427]', '\U00010428', re.I|re.A))
1233+
self.assertTrue(re.match(r'[\U00010400-\U00010427]', '\U00010400', re.I|re.A))
1234+
1235+
self.assertTrue(re.match(r'[N-\x7f]', 'A', re.I|re.A))
1236+
self.assertTrue(re.match(r'[n-\x7f]', 'Z', re.I|re.A))
1237+
self.assertTrue(re.match(r'[N-\uffff]', 'A', re.I|re.A))
1238+
self.assertTrue(re.match(r'[n-\uffff]', 'Z', re.I|re.A))
1239+
self.assertTrue(re.match(r'[N-\U00010000]', 'A', re.I|re.A))
1240+
self.assertTrue(re.match(r'[n-\U00010000]', 'Z', re.I|re.A))
1241+
11871242
# Two different characters have the same lowercase.
11881243
assert 'K'.lower() == '\u212a'.lower() == 'k' # 'K'
11891244
self.assertTrue(re.match(r'[J-M]', '\u212a', re.I))

‎Misc/NEWS.d/next/Library/2024-11-07-22-41-47.gh-issue-126505.iztYE1.rst‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,4 @@
1+
Fix bugs in compiling case-insensitive :mod:`regular expressions <re>` with
2+
character classes containing non-BMP characters: upper-case non-BMP
3+
character did was ignored and the ASCII flag was ignored when
4+
matching a character range whose upper bound is beyond the BMP region.

0 commit comments

Comments
 (0)