diff -r 35da5d848ffd Lib/encodings/idna.py
--- a/Lib/encodings/idna.py	Mon Sep 23 23:24:38 2013 +0300
+++ b/Lib/encodings/idna.py	Tue Sep 24 15:08:59 2013 +0000
@@ -88,7 +88,7 @@
         raise UnicodeError("label empty or too long")
 
     # Step 5: Check ACE prefix
-    if label.startswith(sace_prefix):
+    if label.lower().startswith(sace_prefix):
         raise UnicodeError("Label starts with ACE prefix")
 
     # Step 6: Encode with PUNYCODE
@@ -121,7 +121,7 @@
         except UnicodeError:
             raise UnicodeError("Invalid character in IDN label")
     # Step 3: Check for ACE prefix
-    if not label.startswith(ace_prefix):
+    if not label.lower().startswith(ace_prefix):
         return str(label, "ascii")
 
     # Step 4: Remove ACE prefix
@@ -194,7 +194,7 @@
             # XXX obviously wrong, see #3232
             input = bytes(input)
 
-        if ace_prefix not in input:
+        if ace_prefix not in input.lower():
             # Fast path
             try:
                 return input.decode('ascii'), len(input)
diff -r 35da5d848ffd Lib/test/test_codecs.py
--- a/Lib/test/test_codecs.py	Mon Sep 23 23:24:38 2013 +0300
+++ b/Lib/test/test_codecs.py	Tue Sep 24 15:08:59 2013 +0000
@@ -1371,6 +1371,7 @@
         self.assertEqual(str(b"python.org.", "idna"), "python.org.")
         self.assertEqual(str(b"xn--pythn-mua.org", "idna"), "pyth\xf6n.org")
         self.assertEqual(str(b"xn--pythn-mua.org.", "idna"), "pyth\xf6n.org.")
+        self.assertEqual(str(b"XN--pythn-mua.org.", "idna"), "pyth\xf6n.org.")
 
     def test_builtin_encode(self):
         self.assertEqual("python.org".encode("idna"), b"python.org")