diff --git a/django/template/__init__.py b/django/template/__init__.py
index a592765db6..39ce074852 100644
--- a/django/template/__init__.py
+++ b/django/template/__init__.py
@@ -193,18 +193,27 @@ class Lexer(object):
 
     def tokenize(self):
         "Return a list of tokens from a given template_string"
-        # remove all empty strings, because the regex has a tendency to add them
-        bits = filter(None, tag_re.split(self.template_string))
-        return map(self.create_token, bits)
+        in_tag = False
+        result = []
+        for bit in tag_re.split(self.template_string):
+            if bit:
+                result.append(self.create_token(bit, in_tag))
+            in_tag = not in_tag
+        return result
 
-    def create_token(self,token_string):
-        "Convert the given token string into a new Token object and return it"
-        if token_string.startswith(VARIABLE_TAG_START):
-            token = Token(TOKEN_VAR, token_string[len(VARIABLE_TAG_START):-len(VARIABLE_TAG_END)].strip())
-        elif token_string.startswith(BLOCK_TAG_START):
-            token = Token(TOKEN_BLOCK, token_string[len(BLOCK_TAG_START):-len(BLOCK_TAG_END)].strip())
-        elif token_string.startswith(COMMENT_TAG_START):
-            token = Token(TOKEN_COMMENT, '')
+    def create_token(self, token_string, in_tag=False):
+        """
+        Convert the given token string into a new Token object and return it.
+        If tag is True, we are processing something that matched a tag,
+        otherwise it should be treated as a literal string.
+        """
+        if in_tag:
+            if token_string.startswith(VARIABLE_TAG_START):
+                token = Token(TOKEN_VAR, token_string[len(VARIABLE_TAG_START):-len(VARIABLE_TAG_END)].strip())
+            elif token_string.startswith(BLOCK_TAG_START):
+                token = Token(TOKEN_BLOCK, token_string[len(BLOCK_TAG_START):-len(BLOCK_TAG_END)].strip())
+            elif token_string.startswith(COMMENT_TAG_START):
+                token = Token(TOKEN_COMMENT, '')
         else:
             token = Token(TOKEN_TEXT, token_string)
         return token
@@ -215,21 +224,21 @@ class DebugLexer(Lexer):
 
     def tokenize(self):
         "Return a list of tokens from a given template_string"
-        token_tups, upto = [], 0
+        result, upto = [], 0
         for match in tag_re.finditer(self.template_string):
             start, end = match.span()
             if start > upto:
-                token_tups.append( (self.template_string[upto:start], (upto, start)) )
+                result.append(self.create_token(self.template_string[upto:start], (upto, start), False))
                 upto = start
-            token_tups.append( (self.template_string[start:end], (start,end)) )
+            result.append(self.create_token(self.template_string[start:end], (start, end), True))
             upto = end
         last_bit = self.template_string[upto:]
         if last_bit:
-            token_tups.append( (last_bit, (upto, upto + len(last_bit))) )
-        return [self.create_token(tok, (self.origin, loc)) for tok, loc in token_tups]
+            result.append(self.create_token(last_bit, (upto, upto + len(last_bit)), False))
+        return result
 
-    def create_token(self, token_string, source):
-        token = super(DebugLexer, self).create_token(token_string)
+    def create_token(self, token_string, source, in_tag):
+        token = super(DebugLexer, self).create_token(token_string, in_tag)
         token.source = source
         return token
 
diff --git a/tests/regressiontests/templates/tests.py b/tests/regressiontests/templates/tests.py
index 2b33c53eb1..4e321b2d2b 100644
--- a/tests/regressiontests/templates/tests.py
+++ b/tests/regressiontests/templates/tests.py
@@ -127,6 +127,18 @@ class Templates(unittest.TestCase):
             # Fail silently when accessing a non-simple method
             'basic-syntax20': ("{{ var.method2 }}", {"var": SomeClass()}, ("","INVALID")),
 
+            # Don't get confused when parsing something that is almost, but not
+            # quite, a template tag.
+            'basic-syntax21': ("a {{ moo %} b", {}, "a {{ moo %} b"),
+            'basic-syntax22': ("{{ moo #}", {}, "{{ moo #}"),
+
+            # Will try to treat "moo #} {{ cow" as the variable. Not ideal, but
+            # costly to work around, so this triggers an error.
+            'basic-syntax23': ("{{ moo #} {{ cow }}", {"cow": "cow"}, template.TemplateSyntaxError),
+
+            # Embedded newlines make it not-a-tag.
+            'basic-syntax24': ("{{ moo\n }}", {}, "{{ moo\n }}"),
+
             # List-index syntax allows a template to access a certain item of a subscriptable object.
             'list-index01': ("{{ var.1 }}", {"var": ["first item", "second item"]}, "second item"),