Index: Lib/doctest.py
===================================================================
--- Lib/doctest.py	(revision 80264)
+++ Lib/doctest.py	(working copy)
@@ -263,6 +263,9 @@
         StringIO.truncate(self, size)
         if hasattr(self, "softspace"):
             del self.softspace
+        if not self.buf:
+            # Reset it to an empty string, to make sure it's not unicode:
+            self.buf = ''
 
 # Worst-case linear-time ellipsis matching.
 def _ellipsis_match(want, got):
Index: Lib/test/test_doctest.py
===================================================================
--- Lib/test/test_doctest.py	(revision 80264)
+++ Lib/test/test_doctest.py	(working copy)
@@ -1511,7 +1511,34 @@
     Traceback (most recent call last):
     ValueError: line 0 of the doctest for s has an option directive on a line with no example: '# doctest: +ELLIPSIS'
 """
+    
+    
 
+    def test_unicode_output(self): r"""
+
+Check that unicode output works:
+    
+    >>> u'\xe9'
+    u'\xe9'
+
+If we return unicode, SpoofOut's buf variable becomes automagically
+converted to unicode. This means all subsequent output becomes converted
+to unicode, and if the output contains non-ascii characters that failed.
+It used to be that this state change carried on between tests, meaning
+tests would fail if unicode has been output previously in the testrun.
+This test tests that this ois no longer so:
+    
+    >>> print u'abc'
+    abc
+
+And then return a string with non-ascii characters:
+
+    >>> print u'\xe9'.encode('utf-8')
+    é
+    
+"""
+    
+
 def test_testsource(): r"""
 Unit tests for `testsource()`.