use cleaner, less safe, unicode/str in LazyEvaluate

IPython/core/prompts.py

@@ -95,10 +95,7 b' class LazyEvaluate(object):'
         return self.func(*self.args, **self.kwargs)
     
     def __str__(self):
-        s = self()
-        if isinstance(s, unicode):
-            return py3compat.unicode_to_str(s)
-        return str(s)
+        return str(self())
     
     def __unicode__(self):
         return unicode(self())

IPython/core/tests/test_prompts.py

@@ -77,7 +77,7 b' class PromptTests(unittest.TestCase):'
     def test_lazy_eval_unicode(self):
         u = u'ünicødé'
         lz = LazyEvaluate(lambda : u)
-        self.assertEquals(str(lz), py3compat.unicode_to_str(u))
+        # str(lz) would fail
         self.assertEquals(unicode(lz), u)
         self.assertEquals(format(lz), u)
     
@@ -85,10 +85,9 b' class PromptTests(unittest.TestCase):'
         u = u'ünicødé'
         b = u.encode('utf8')
         lz = LazyEvaluate(lambda : b)
+        # unicode(lz) would fail
         self.assertEquals(str(lz), str(b))
         self.assertEquals(format(lz), str(b))
-        if not py3compat.PY3:
-            self.assertRaises(UnicodeDecodeError, unicode, lz)
     
     def test_lazy_eval_float(self):
         f = 0.503