Issues #814253, #9179: Warnings now are raised when group references and

conditional group references are used in lookbehind assertions in regular
expressions.

diff --git a/Doc/library/re.rst b/Doc/library/re.rst
index dfc25ea08915bfdb88dbc1b5efc4f4ae87b917b0..c3c8b65d8d925af9dd86f37cd60881df1e2b4e0d 100644 (file)
--- a/Doc/library/re.rst
+++ b/Doc/library/re.rst
@@ -281,7 +281,9 @@ The special characters are:
    assertion`. ``(?<=abc)def`` will find a match in ``abcdef``, since the
    lookbehind will back up 3 characters and check if the contained pattern matches.
    The contained pattern must only match strings of some fixed length, meaning that
-   ``abc`` or ``a|b`` are allowed, but ``a*`` and ``a{3,4}`` are not.  Note that
+   ``abc`` or ``a|b`` are allowed, but ``a*`` and ``a{3,4}`` are not.  Group
+   references are not supported even if they match strings of some fixed length.
+   Note that
    patterns which start with positive lookbehind assertions will not match at the
    beginning of the string being searched; you will most likely want to use the
    :func:`search` function rather than the :func:`match` function:
@@ -301,7 +303,8 @@ The special characters are:
    Matches if the current position in the string is not preceded by a match for
    ``...``.  This is called a :dfn:`negative lookbehind assertion`.  Similar to
    positive lookbehind assertions, the contained pattern must only match strings of
-   some fixed length.  Patterns which start with negative lookbehind assertions may
+   some fixed length and shouldn't contain group references.
+   Patterns which start with negative lookbehind assertions may
    match at the beginning of the string being searched.
 
 ``(?(id/name)yes-pattern|no-pattern)``

diff --git a/Lib/sre_parse.py b/Lib/sre_parse.py
index b56d437c3479d04fc1f0351071bad85c9d8d7be1..df1e6437c7dbb40d6c495016a8736f968b28f0b1 100644 (file)
--- a/Lib/sre_parse.py
+++ b/Lib/sre_parse.py
@@ -69,6 +69,8 @@ class Pattern:
         self.open = []
         self.groups = 1
         self.groupdict = {}
+        self.lookbehind = 0
+
     def opengroup(self, name=None):
         gid = self.groups
         self.groups = gid + 1
@@ -352,6 +354,11 @@ def _escape(source, escape, state):
             if group < state.groups:
                 if not state.checkgroup(group):
                     raise error("cannot refer to open group")
+                if state.lookbehind:
+                    import warnings
+                    warnings.warn('group references in lookbehind '
+                                  'assertions are not supported',
+                                  RuntimeWarning)
                 return GROUPREF, group
             raise ValueError
         if len(escape) == 2:
@@ -630,6 +637,11 @@ def _parse(source, state):
                         if gid is None:
                             msg = "unknown group name: {0!r}".format(name)
                             raise error(msg)
+                        if state.lookbehind:
+                            import warnings
+                            warnings.warn('group references in lookbehind '
+                                          'assertions are not supported',
+                                          RuntimeWarning)
                         subpatternappend((GROUPREF, gid))
                         continue
                     else:
@@ -658,7 +670,10 @@ def _parse(source, state):
                             raise error("syntax error")
                         dir = -1 # lookbehind
                         char = sourceget()
+                        state.lookbehind += 1
                     p = _parse_sub(source, state)
+                    if dir < 0:
+                        state.lookbehind -= 1
                     if not sourcematch(")"):
                         raise error("unbalanced parenthesis")
                     if char == "=":
@@ -689,6 +704,11 @@ def _parse(source, state):
                             condgroup = int(condname)
                         except ValueError:
                             raise error("bad character in group name")
+                    if state.lookbehind:
+                        import warnings
+                        warnings.warn('group references in lookbehind '
+                                      'assertions are not supported',
+                                      RuntimeWarning)
                 else:
                     # flags
                     if not source.next in FLAGS:

diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py
index d2547d4a96c62745c86550a34a0069aeca07f22e..7348af3f1aaeb32032cacaf9369f16dae1225477 100644 (file)
--- a/Lib/test/test_re.py
+++ b/Lib/test/test_re.py
@@ -557,7 +557,7 @@ class ReTests(unittest.TestCase):
         self.assertEqual(re.match("a.*b", "a\n\nb", re.DOTALL).group(0),
                          "a\n\nb")
 
-    def test_non_consuming(self):
+    def test_lookahead(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")
@@ -571,6 +571,37 @@ 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'))
+        # Named group reference.
+        self.assertTrue(re.match(r'(?P<g>a)b(?=(?P=g))a', 'aba'))
+        self.assertIsNone(re.match(r'(?P<g>a)b(?=(?P=g))c', 'abac'))
+        # Conditional group reference.
+        self.assertTrue(re.match(r'(?:(a)|(x))b(?=(?(2)x|c))c', 'abc'))
+        self.assertIsNone(re.match(r'(?:(a)|(x))b(?=(?(2)c|x))c', 'abc'))
+        self.assertTrue(re.match(r'(?:(a)|(x))b(?=(?(2)x|c))c', 'abc'))
+        self.assertIsNone(re.match(r'(?:(a)|(x))b(?=(?(1)b|x))c', 'abc'))
+        self.assertTrue(re.match(r'(?:(a)|(x))b(?=(?(1)c|x))c', 'abc'))
+        # Group used before defined.
+        self.assertTrue(re.match(r'(a)b(?=(?(2)x|c))(c)', 'abc'))
+        self.assertIsNone(re.match(r'(a)b(?=(?(2)b|x))(c)', 'abc'))
+        self.assertTrue(re.match(r'(a)b(?=(?(1)c|x))(c)', 'abc'))
+
+    def test_lookbehind(self):
+        self.assertTrue(re.match(r'ab(?<=b)c', 'abc'))
+        self.assertIsNone(re.match(r'ab(?<=c)c', 'abc'))
+        self.assertIsNone(re.match(r'ab(?<!b)c', 'abc'))
+        self.assertTrue(re.match(r'ab(?<!c)c', 'abc'))
+        # Group reference.
+        self.assertWarns(RuntimeWarning, re.compile, r'(a)a(?<=\1)c')
+        # Named group reference.
+        self.assertWarns(RuntimeWarning, re.compile, r'(?P<g>a)a(?<=(?P=g))c')
+        # Conditional group reference.
+        self.assertWarns(RuntimeWarning, re.compile, r'(a)b(?<=(?(1)b|x))c')
+        # Group used before defined.
+        self.assertWarns(RuntimeWarning, re.compile, r'(a)b(?<=(?(2)b|x))(c)')
+
     def test_ignore_case(self):
         self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
         self.assertEqual(re.match(b"abc", b"ABC", re.I).group(0), b"ABC")

diff --git a/Misc/NEWS b/Misc/NEWS
index f1426d2f728ac9b24eb835f0a6275245c98a60fa..37ff72ea1338bdf418e0abdde23b7d754bc67048 100644 (file)
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -13,6 +13,10 @@ Core and Builtins
 Library
 -------
 
+- Issues #814253, #9179: Warnings now are raised when group references and
+  conditional group references are used in lookbehind assertions in regular
+  expressions.
+
 - Issue #23215: Multibyte codecs with custom error handlers that ignores errors
   consumed too much memory and raised SystemError or MemoryError.
   Original patch by Aleksi Torhamo.