#12759: sre_parse now raises a proper error when the name of the group is missing. Initial patch by Serhiy Storchaka.

Lib/sre_parse.py

@@ -548,6 +548,8 @@ def _parse(source, state):
                                 break
                             name = name + char
                         group = 1
+                        if not name:
+                            raise error("missing group name")
                         if not isname(name):
                             raise error("bad character in group name")
                     elif sourcematch("="):
@@ -560,6 +562,8 @@ def _parse(source, state):
                             if char == ")":
                                 break
                             name = name + char
+                        if not name:
+                            raise error("missing group name")
                         if not isname(name):
                             raise error("bad character in group name")
                         gid = state.groupdict.get(name)
@@ -612,6 +616,8 @@ def _parse(source, state):
                             break
                         condname = condname + char
                     group = 2
+                    if not condname:
+                        raise error("missing group name")
                     if isname(condname):
                         condgroup = state.groupdict.get(condname)
                         if condgroup is None:
@@ -743,7 +749,7 @@ def parse_template(source, pattern):
                             break
                         name = name + char
                 if not name:
-                    raise error("bad group name")
+                    raise error("missing group name")
                 try:
                     index = int(name)
                     if index < 0:

Lib/test/test_re.py

@@ -161,11 +161,31 @@ class ReTests(unittest.TestCase):
         self.assertEqual(re.sub('x*', '-', 'abxd'), '-a-b-d-')
         self.assertEqual(re.sub('x+', '-', 'abxd'), 'ab-d')
 
+    def test_symbolic_groups(self):
+        re.compile('(?P<a>x)(?P=a)(?(a)y)')
+        re.compile('(?P<a1>x)(?P=a1)(?(a1)y)')
+        self.assertRaises(re.error, re.compile, '(?P<a>)(?P<a>)')
+        self.assertRaises(re.error, re.compile, '(?Px)')
+        self.assertRaises(re.error, re.compile, '(?P=)')
+        self.assertRaises(re.error, re.compile, '(?P=1)')
+        self.assertRaises(re.error, re.compile, '(?P=a)')
+        self.assertRaises(re.error, re.compile, '(?P=a1)')
+        self.assertRaises(re.error, re.compile, '(?P=a.)')
+        self.assertRaises(re.error, re.compile, '(?P<)')
+        self.assertRaises(re.error, re.compile, '(?P<>)')
+        self.assertRaises(re.error, re.compile, '(?P<1>)')
+        self.assertRaises(re.error, re.compile, '(?P<a.>)')
+        self.assertRaises(re.error, re.compile, '(?())')
+        self.assertRaises(re.error, re.compile, '(?(a))')
+        self.assertRaises(re.error, re.compile, '(?(1a))')
+        self.assertRaises(re.error, re.compile, '(?(a.))')
+
     def test_symbolic_refs(self):
         self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<a', 'xx')
         self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<', 'xx')
         self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g', 'xx')
         self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<a a>', 'xx')
+        self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<>', 'xx')
         self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<1a1>', 'xx')
         self.assertRaises(IndexError, re.sub, '(?P<a>x)', '\g<ab>', 'xx')
         self.assertRaises(re.error, re.sub, '(?P<a>x)|(?P<b>y)', '\g<b>', 'xx')

Misc/NEWS

@@ -143,6 +143,9 @@ Core and Builtins
 Library
 -------
 
+- Issue #12759: sre_parse now raises a proper error when the name of the group
+  is missing.  Initial patch by Serhiy Storchaka.
+
 - Issue #16152: fix tokenize to ignore whitespace at the end of the code when
   no newline is found.  Patch by Ned Batchelder.