Fix #1528 -- Yaml gets confused when a comment contains a key:value pair.

3 files changed, 20 insertions, 2 deletions

diff --git a/CHANGES b/CHANGES
index eac1a3bb..ac233630 100644
--- a/CHANGES
+++ b/CHANGES
@@ -13,6 +13,7 @@ Version 2.4.2
 - Updated lexers:
 
   * Erlang, Elxir shells (PR#823, #1521)
+  * YAML (#1528)
 
 - Fix encoding error when guessing lexer with given ``encoding`` option
   (#1438)
diff --git a/pygments/lexers/data.py b/pygments/lexers/data.py
index b3253542..46ca7340 100644
--- a/pygments/lexers/data.py
+++ b/pygments/lexers/data.py
@@ -233,7 +233,7 @@ class YamlLexer(ExtendedRegexLexer):
             # whitespaces separating tokens
             (r'[ ]+', Text),
             # key with colon
-            (r'''([^,:?\[\]{}"'\n]+)(:)(?=[ ]|$)''',
+            (r'''([^#,:?\[\]{}"'\n]+)(:)(?=[ ]|$)''',
              bygroups(Name.Tag, set_indent(Punctuation, implicit=True))),
             # tags, anchors and aliases,
             include('descriptors'),
diff --git a/tests/test_data.py b/tests/test_data.py
index be371419..20e74be0 100644
--- a/tests/test_data.py
+++ b/tests/test_data.py
@@ -9,7 +9,7 @@
 
 import unittest
 
-from pygments.lexers import JsonLexer, JsonBareObjectLexer
+from pygments.lexers import JsonLexer, JsonBareObjectLexer, YamlLexer
 from pygments.token import Token
 
 
@@ -98,3 +98,20 @@ class JsonBareObjectTest(unittest.TestCase):
         ]
         self.assertEqual(tokens, list(self.lexer.get_tokens(fragment)))
 
+class YamlTest(unittest.TestCase):
+    def setUp(self):
+        self.lexer = YamlLexer()
+
+    def testColonInComment(self):
+        # Bug #1528: This previously parsed 'token # innocent' as a tag
+        fragment = u'here: token # innocent: comment\n'
+        tokens = [
+            (Token.Name.Tag, u'here'),
+            (Token.Punctuation, u':'),
+            (Token.Text, u' '),
+            (Token.Literal.Scalar.Plain, u'token'),
+            (Token.Text, u' '),
+            (Token.Comment.Single, u'# innocent: comment'),
+            (Token.Text, u'\n'),
+        ]
+        self.assertEqual(tokens, list(self.lexer.get_tokens(fragment)))