diff --git a/Lib/re.py b/Lib/re.py index 788fa6bd789..bc3056fdbfa 100644 --- a/Lib/re.py +++ b/Lib/re.py @@ -351,11 +351,10 @@ class Scanner: s = sre_parse.Pattern() s.flags = flags for phrase, action in lexicon: - gid = s.opengroup() p.append(sre_parse.SubPattern(s, [ - (SUBPATTERN, (gid, sre_parse.parse(phrase, flags))), + (SUBPATTERN, (len(p)+1, sre_parse.parse(phrase, flags))), ])) - s.closegroup(gid, p[-1]) + s.groups = len(p)+1 p = sre_parse.SubPattern(s, [(BRANCH, (None, p))]) self.scanner = sre_compile.compile(p) def scan(self, string): diff --git a/Lib/sre_parse.py b/Lib/sre_parse.py index 45411f89f16..8d9a09a860d 100644 --- a/Lib/sre_parse.py +++ b/Lib/sre_parse.py @@ -68,14 +68,12 @@ class Pattern: # master pattern object. keeps track of global attributes def __init__(self): self.flags = 0 + self.open = [] + self.groups = 1 self.groupdict = {} - self.subpatterns = [None] # group 0 - @property - def groups(self): - return len(self.subpatterns) def opengroup(self, name=None): gid = self.groups - self.subpatterns.append(None) + self.groups = gid + 1 if self.groups > MAXGROUPS: raise error("groups number is too large") if name is not None: @@ -84,11 +82,12 @@ class Pattern: raise error("redefinition of group name %r as group %d; " "was group %d" % (name, gid, ogid)) self.groupdict[name] = gid + self.open.append(gid) return gid - def closegroup(self, gid, p): - self.subpatterns[gid] = p + def closegroup(self, gid): + self.open.remove(gid) def checkgroup(self, gid): - return gid < self.groups and self.subpatterns[gid] is not None + return gid < self.groups and gid not in self.open class SubPattern: # a subpattern, in intermediate form @@ -184,21 +183,7 @@ class SubPattern: elif op in _UNITCODES: lo = lo + 1 hi = hi + 1 - elif op is GROUPREF: - i, j = self.pattern.subpatterns[av].getwidth() - lo = lo + i - hi = hi + j - elif op is GROUPREF_EXISTS: - i, j = av[1].getwidth() - if av[2] is not None: - l, h = av[2].getwidth() - i = min(i, l) - j = max(j, h) - else: - i = 0 - lo = lo + i - hi = hi + j - elif op is SUCCESS: + elif op == SUCCESS: break self.width = min(lo, MAXREPEAT - 1), min(hi, MAXREPEAT) return self.width @@ -741,7 +726,7 @@ def _parse(source, state): if not sourcematch(")"): raise source.error("unbalanced parenthesis") if group is not None: - state.closegroup(group, p) + state.closegroup(group) subpatternappend((SUBPATTERN, (group, p))) else: while True: diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py index 7bc1e935d46..83b1425d989 100644 --- a/Lib/test/test_re.py +++ b/Lib/test/test_re.py @@ -589,7 +589,7 @@ class ReTests(unittest.TestCase): self.assertEqual(re.match("a.*b", "a\n\nb", re.DOTALL).group(0), "a\n\nb") - def test_lookahead(self): + def test_non_consuming(self): self.assertEqual(re.match("(a(?=\s[^a]))", "a b").group(1), "a") self.assertEqual(re.match("(a(?=\s[^a]*))", "a b").group(1), "a") self.assertEqual(re.match("(a(?=\s[abc]))", "a b").group(1), "a") @@ -603,42 +603,6 @@ class ReTests(unittest.TestCase): self.assertEqual(re.match(r"(a)(?!\s\1)", "a b").group(1), "a") self.assertEqual(re.match(r"(a)(?!\s(abc|a))", "a b").group(1), "a") - # Group reference. - self.assertTrue(re.match(r'(a)b(?=\1)a', 'aba')) - self.assertIsNone(re.match(r'(a)b(?=\1)c', 'abac')) - # Conditional group reference. - self.assertTrue(re.match('(?:(a)|(x))b(?=(?(2)x|c))c', 'abc')) - self.assertIsNone(re.match('(?:(a)|(x))b(?=(?(2)c|x))c', 'abc')) - self.assertTrue(re.match('(?:(a)|(x))b(?=(?(2)x|c))c', 'abc')) - self.assertIsNone(re.match('(?:(a)|(x))b(?=(?(1)b|x))c', 'abc')) - self.assertTrue(re.match('(?:(a)|(x))b(?=(?(1)c|x))c', 'abc')) - # Group used before defined. - self.assertTrue(re.match('(a)b(?=(?(2)x|c))(c)', 'abc')) - self.assertIsNone(re.match('(a)b(?=(?(2)b|x))(c)', 'abc')) - self.assertTrue(re.match('(a)b(?=(?(1)c|x))(c)', 'abc')) - - def test_lookbehind(self): - self.assertTrue(re.match('ab(?<=b)c', 'abc')) - self.assertIsNone(re.match('ab(?<=c)c', 'abc')) - self.assertIsNone(re.match('ab(?