Revert "bpo-23689: re module, fix memory leak when a match is terminated by a signal or memory allocation failure (pythonGH-32283)"

This reverts commit 6e3eee5.

Manual fixups to increase the MAGIC number and to handle conflicts with
a couple of changes that landed after that.

Author: gpshead

Lib/re/_compiler.py

@@ -28,21 +28,54 @@
     POSSESSIVE_REPEAT: (POSSESSIVE_REPEAT, SUCCESS, POSSESSIVE_REPEAT_ONE),
 }
 
-class _CompileData:
-    __slots__ = ('code', 'repeat_count')
-    def __init__(self):
-        self.code = []
-        self.repeat_count = 0
+# 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 _combine_flags(flags, add_flags, del_flags,
                    TYPE_FLAGS=_parser.TYPE_FLAGS):
     if add_flags & TYPE_FLAGS:
         flags &= ~TYPE_FLAGS
     return (flags | add_flags) & ~del_flags
 
-def _compile(data, pattern, flags):
+def _compile(code, pattern, flags):
     # internal: compile a (sub)pattern
-    code = data.code
     emit = code.append
     _len = len
     LITERAL_CODES = _LITERAL_CODES
@@ -115,19 +148,15 @@ def _compile(data, pattern, flags):
                 skip = _len(code); emit(0)
                 emit(av[0])
                 emit(av[1])
-                _compile(data, av[2], flags)
+                _compile(code, av[2], flags)
                 emit(SUCCESS)
                 code[skip] = _len(code) - skip
             else:
                 emit(REPEATING_CODES[op][0])
                 skip = _len(code); emit(0)
                 emit(av[0])
                 emit(av[1])
-                # now op is in (MIN_REPEAT, MAX_REPEAT, POSSESSIVE_REPEAT)
-                if op != POSSESSIVE_REPEAT:
-                    emit(data.repeat_count)
-                    data.repeat_count += 1
-                _compile(data, av[2], flags)
+                _compile(code, av[2], flags)
                 code[skip] = _len(code) - skip
                 emit(REPEATING_CODES[op][1])
         elif op is SUBPATTERN:
@@ -136,7 +165,7 @@ def _compile(data, pattern, flags):
                 emit(MARK)
                 emit((group-1)*2)
             # _compile_info(code, p, _combine_flags(flags, add_flags, del_flags))
-            _compile(data, p, _combine_flags(flags, add_flags, del_flags))
+            _compile(code, p, _combine_flags(flags, add_flags, del_flags))
             if group:
                 emit(MARK)
                 emit((group-1)*2+1)
@@ -148,7 +177,7 @@ def _compile(data, pattern, flags):
             # pop their stack if they reach it
             emit(ATOMIC_GROUP)
             skip = _len(code); emit(0)
-            _compile(data, av, flags)
+            _compile(code, av, flags)
             emit(SUCCESS)
             code[skip] = _len(code) - skip
         elif op in SUCCESS_CODES:
@@ -163,7 +192,7 @@ def _compile(data, pattern, flags):
                 if lo != hi:
                     raise error("look-behind requires fixed-width pattern")
                 emit(lo) # look behind
-            _compile(data, av[1], flags)
+            _compile(code, av[1], flags)
             emit(SUCCESS)
             code[skip] = _len(code) - skip
         elif op is AT:
@@ -182,7 +211,7 @@ def _compile(data, pattern, flags):
             for av in av[1]:
                 skip = _len(code); emit(0)
                 # _compile_info(code, av, flags)
-                _compile(data, av, flags)
+                _compile(code, av, flags)
                 emit(JUMP)
                 tailappend(_len(code)); emit(0)
                 code[skip] = _len(code) - skip
@@ -210,12 +239,12 @@ def _compile(data, pattern, flags):
             emit(op)
             emit(av[0]-1)
             skipyes = _len(code); emit(0)
-            _compile(data, av[1], flags)
+            _compile(code, av[1], flags)
             if av[2]:
                 emit(JUMP)
                 skipno = _len(code); emit(0)
                 code[skipyes] = _len(code) - skipyes + 1
-                _compile(data, av[2], flags)
+                _compile(code, av[2], flags)
                 code[skipno] = _len(code) - skipno
             else:
                 code[skipyes] = _len(code) - skipyes + 1
@@ -582,17 +611,17 @@ def isstring(obj):
 def _code(p, flags):
 
     flags = p.state.flags | flags
-    data = _CompileData()
+    code = []
 
     # compile info block
-    _compile_info(data.code, p, flags)
+    _compile_info(code, p, flags)
 
     # compile the pattern
-    _compile(data, p.data, flags)
+    _compile(code, p.data, flags)
 
-    data.code.append(SUCCESS)
+    code.append(SUCCESS)
 
-    return data
+    return code
 
 def _hex_code(code):
     return '[%s]' % ', '.join('%#0*x' % (_sre.CODESIZE*2+2, x) for x in code)
@@ -693,21 +722,14 @@ def print_2(*args):
                     else:
                         print_(FAILURE)
                 i += 1
-            elif op in (REPEAT_ONE, MIN_REPEAT_ONE,
+            elif op in (REPEAT, REPEAT_ONE, MIN_REPEAT_ONE,
                         POSSESSIVE_REPEAT, POSSESSIVE_REPEAT_ONE):
                 skip, min, max = code[i: i+3]
                 if max == MAXREPEAT:
                     max = 'MAXREPEAT'
                 print_(op, skip, min, max, to=i+skip)
                 dis_(i+3, i+skip)
                 i += skip
-            elif op is REPEAT:
-                skip, min, max, repeat_index = code[i: i+4]
-                if max == MAXREPEAT:
-                    max = 'MAXREPEAT'
-                print_(op, skip, min, max, repeat_index, to=i+skip)
-                dis_(i+4, i+skip)
-                i += skip
             elif op is GROUPREF_EXISTS:
                 arg, skip = code[i: i+2]
                 print_(op, arg, skip, to=i+skip)
@@ -762,11 +784,11 @@ def compile(p, flags=0):
     else:
         pattern = None
 
-    data = _code(p, flags)
+    code = _code(p, flags)
 
     if flags & SRE_FLAG_DEBUG:
         print()
-        dis(data.code)
+        dis(code)
 
     # map in either direction
     groupindex = p.state.groupdict
@@ -775,6 +797,7 @@ def compile(p, flags=0):
         indexgroup[i] = k
 
     return _sre.compile(
-        pattern, flags | p.state.flags, data.code,
-        p.state.groups-1, groupindex, tuple(indexgroup),
-        data.repeat_count)
+        pattern, flags | p.state.flags, code,
+        p.state.groups-1,
+        groupindex, tuple(indexgroup)
+        )

Lib/re/_constants.py

@@ -13,7 +13,7 @@
 
 # update when constants are added or removed
 
-MAGIC = 20220423
+MAGIC = 20220615
 
 from _sre import MAXREPEAT, MAXGROUPS
 

Lib/test/test_re.py

@@ -1765,12 +1765,9 @@ def test_dealloc(self):
         long_overflow = 2**128
         self.assertRaises(TypeError, re.finditer, "a", {})
         with self.assertRaises(OverflowError):
-            _sre.compile("abc", 0, [long_overflow], 0, {}, (), 0)
+            _sre.compile("abc", 0, [long_overflow], 0, {}, ())
         with self.assertRaises(TypeError):
-            _sre.compile({}, 0, [], 0, [], [], 0)
-        with self.assertRaises(RuntimeError):
-            # invalid repeat_count -1
-            _sre.compile("abc", 0, [1], 0, {}, (), -1)
+            _sre.compile({}, 0, [], 0, [], [])
 
     def test_search_dot_unicode(self):
         self.assertTrue(re.search("123.*-", '123abc-'))
@@ -2509,27 +2506,6 @@ def test_possesive_repeat(self):
 14. SUCCESS
 ''')
 
-    def test_repeat_index(self):
-        self.assertEqual(get_debug_out(r'(?:ab)*?(?:cd)*'), '''\
-MIN_REPEAT 0 MAXREPEAT
-  LITERAL 97
-  LITERAL 98
-MAX_REPEAT 0 MAXREPEAT
-  LITERAL 99
-  LITERAL 100
-
- 0. INFO 4 0b0 0 MAXREPEAT (to 5)
- 5: REPEAT 8 0 MAXREPEAT 0 (to 14)
-10.   LITERAL 0x61 ('a')
-12.   LITERAL 0x62 ('b')
-14: MIN_UNTIL
-15. REPEAT 8 0 MAXREPEAT 1 (to 24)
-20.   LITERAL 0x63 ('c')
-22.   LITERAL 0x64 ('d')
-24: MAX_UNTIL
-25. SUCCESS
-''')
-
 
 class PatternReprTests(unittest.TestCase):
     def check(self, pattern, expected):