gh-91575: Update case-insensitive matching in re to the latest Unicode version

Lib/re/_compiler.py

@@ -57,6 +57,22 @@
     (0x3c2, 0x3c3), # ςσ
     # GREEK SMALL LETTER PHI, GREEK PHI SYMBOL
     (0x3c6, 0x3d5), # φϕ
+    # CYRILLIC SMALL LETTER VE, CYRILLIC SMALL LETTER ROUNDED VE
+    (0x432, 0x1c80), # вᲀ
+    # CYRILLIC SMALL LETTER DE, CYRILLIC SMALL LETTER LONG-LEGGED DE
+    (0x434, 0x1c81), # дᲁ
+    # CYRILLIC SMALL LETTER O, CYRILLIC SMALL LETTER NARROW O
+    (0x43e, 0x1c82), # оᲂ
+    # CYRILLIC SMALL LETTER ES, CYRILLIC SMALL LETTER WIDE ES
+    (0x441, 0x1c83), # сᲃ
+    # CYRILLIC SMALL LETTER TE, CYRILLIC SMALL LETTER TALL TE, CYRILLIC SMALL LETTER THREE-LEGGED TE
+    (0x442, 0x1c84, 0x1c85), # тᲄᲅ
+    # CYRILLIC SMALL LETTER HARD SIGN, CYRILLIC SMALL LETTER TALL HARD SIGN
+    (0x44a, 0x1c86), # ъᲆ
+    # CYRILLIC SMALL LETTER YAT, CYRILLIC SMALL LETTER TALL YAT
+    (0x463, 0x1c87), # ѣᲇ
+    # CYRILLIC SMALL LETTER UNBLENDED UK, CYRILLIC SMALL LETTER MONOGRAPH UK
+    (0x1c88, 0xa64b), # ᲈꙋ
     # 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
@@ -339,11 +355,19 @@ def _optimize_charset(charset, iscased=None, fixup=None, fixes=None):
                     charmap += b'\0' * 0xff00
                     continue
                 # Character set contains non-BMP character codes.
+                # For range, all BMP characters in the range are already
+                # proceeded.
                 if fixup:
                     hascased = True
-                    # There are only two ranges of cased non-BMP characters:
-                    # 10400-1044F (Deseret) and 118A0-118DF (Warang Citi),
-                    # and for both ranges RANGE_UNI_IGNORE works.
+                    # For now, IN_UNI_IGNORE+LITERAL and
+                    # IN_UNI_IGNORE+RANGE_UNI_IGNORE work for all non-BMP
+                    # characters, because two characters (at least one of
+                    # which is not in the BMP) match case-insensitively
+                    # if and only if:
+                    # 1) c1.lower() == c2.lower()
+                    # 2) c1.lower() == c2 or c1.lower().upper() == c2
+                    # Also, both c.lower() and c.lower().upper() are single
+                    # characters for every non-BMP character.
                     if op is RANGE:
                         op = RANGE_UNI_IGNORE
                 tail.append((op, av))

Lib/test/test_re.py

@@ -859,16 +859,30 @@ def test_ignore_case(self):
         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'
+        # Two different characters have the same lowercase.
+        assert 'K'.lower() == '\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'
+
+        # Two different characters have the same uppercase.
+        assert 's'.upper() == '\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))
+
+        # Two different characters have the same uppercase. Unicode 9.0+.
+        assert '\u0432'.upper() == '\u1c80'.upper() == '\u0412' # 'в', 'ᲀ', 'В'
+        self.assertTrue(re.match(r'\u0412', '\u0432', re.I))
+        self.assertTrue(re.match(r'\u0412', '\u1c80', re.I))
+        self.assertTrue(re.match(r'\u0432', '\u0412', re.I))
+        self.assertTrue(re.match(r'\u0432', '\u1c80', re.I))
+        self.assertTrue(re.match(r'\u1c80', '\u0412', re.I))
+        self.assertTrue(re.match(r'\u1c80', '\u0432', re.I))
+
+        # Two different characters have the same multicharacter uppercase.
         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))
@@ -882,16 +896,31 @@ def test_ignore_case_set(self):
         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'
+
+        # Two different characters have the same lowercase.
+        assert 'K'.lower() == '\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'
+
+        # Two different characters have the same uppercase.
+        assert 's'.upper() == '\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))
+
+        # Two different characters have the same uppercase. Unicode 9.0+.
+        assert '\u0432'.upper() == '\u1c80'.upper() == '\u0412' # 'в', 'ᲀ', 'В'
+        self.assertTrue(re.match(r'[19\u0412]', '\u0432', re.I))
+        self.assertTrue(re.match(r'[19\u0412]', '\u1c80', re.I))
+        self.assertTrue(re.match(r'[19\u0432]', '\u0412', re.I))
+        self.assertTrue(re.match(r'[19\u0432]', '\u1c80', re.I))
+        self.assertTrue(re.match(r'[19\u1c80]', '\u0412', re.I))
+        self.assertTrue(re.match(r'[19\u1c80]', '\u0432', re.I))
+
+        # Two different characters have the same multicharacter uppercase.
         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))
@@ -915,16 +944,30 @@ def test_ignore_case_range(self):
         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'
+        # Two different characters have the same lowercase.
+        assert 'K'.lower() == '\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'
+
+        # Two different characters have the same uppercase.
+        assert 's'.upper() == '\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))
+
+        # Two different characters have the same uppercase. Unicode 9.0+.
+        assert '\u0432'.upper() == '\u1c80'.upper() == '\u0412' # 'в', 'ᲀ', 'В'
+        self.assertTrue(re.match(r'[\u0411-\u0413]', '\u0432', re.I))
+        self.assertTrue(re.match(r'[\u0411-\u0413]', '\u1c80', re.I))
+        self.assertTrue(re.match(r'[\u0431-\u0433]', '\u0412', re.I))
+        self.assertTrue(re.match(r'[\u0431-\u0433]', '\u1c80', re.I))
+        self.assertTrue(re.match(r'[\u1c80-\u1c82]', '\u0412', re.I))
+        self.assertTrue(re.match(r'[\u1c80-\u1c82]', '\u0432', re.I))
+
+        # Two different characters have the same multicharacter uppercase.
         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))

Misc/NEWS.d/next/Library/2022-04-15-18-38-21.gh-issue-91575.fSyAxS.rst

@@ -0,0 +1,2 @@
+Update case-insensitive matching in the :mod:`re` module to the latest
+Unicode version.