Fix a bug in default grouping in regexopt when the words are all the same length.

Fixes #1229

2 files changed, 35 insertions, 1 deletions

diff --git a/pygments/regexopt.py b/pygments/regexopt.py
index 79903684..047c703f 100644
--- a/pygments/regexopt.py
+++ b/pygments/regexopt.py
@@ -54,7 +54,7 @@ def regex_opt_inner(strings, open_paren):
                 return open_paren + regex_opt_inner(rest, '') + '|' \
                     + make_charset(oneletter) + close_paren
             # print '-> only 1-character'
-            return make_charset(oneletter)
+            return open_paren + make_charset(oneletter) + close_paren
     prefix = commonprefix(strings)
     if prefix:
         plen = len(prefix)
diff --git a/tests/test_regexopt.py b/tests/test_regexopt.py
index dd56a446..6322c735 100644
--- a/tests/test_regexopt.py
+++ b/tests/test_regexopt.py
@@ -46,6 +46,7 @@ class RegexOptTestCase(unittest.TestCase):
                                      random.randint(1, len(kwlist) - 1))
             no_match = set(kwlist) - set(to_match)
             rex = re.compile(regex_opt(to_match))
+            self.assertEqual(rex.groups, 1)
             for w in to_match:
                 self.assertTrue(rex.match(w))
             for w in no_match:
@@ -74,3 +75,36 @@ class RegexOptTestCase(unittest.TestCase):
         rex = re.compile(opt)
         m = rex.match('abfoo')
         self.assertEqual(5, m.end())
+
+    def test_different_length_grouping(self):
+        opt = regex_opt(('a', 'xyz'))
+        print(opt)
+        rex = re.compile(opt)
+        self.assertTrue(rex.match('a'))
+        self.assertTrue(rex.match('xyz'))
+        self.assertFalse(rex.match('b'))
+        self.assertEqual(1, rex.groups)
+
+    def test_same_length_grouping(self):
+        opt = regex_opt(('a', 'b'))
+        print(opt)
+        rex = re.compile(opt)
+        self.assertTrue(rex.match('a'))
+        self.assertTrue(rex.match('b'))
+        self.assertFalse(rex.match('x'))
+
+        self.assertEqual(1, rex.groups)
+        groups = rex.match('a').groups()
+        self.assertEqual(('a',), groups)
+
+    def test_same_length_suffix_grouping(self):
+        opt = regex_opt(('a', 'b'), suffix='(m)')
+        print(opt)
+        rex = re.compile(opt)
+        self.assertTrue(rex.match('am'))
+        self.assertTrue(rex.match('bm'))
+        self.assertFalse(rex.match('xm'))
+        self.assertFalse(rex.match('ax'))
+        self.assertEqual(2, rex.groups)
+        groups = rex.match('am').groups()
+        self.assertEqual(('a', 'm'), groups)