Decode UTF-8 with 'replace' option to avoid exceptions

To avoid generating an exception when receiving and trying to decode non-
UTF-8 encoded text, use the decode 'replace' option -- this replaces all
malformed encodings with uFFFD REPLACEMENT CHARACTER.

This addresses issue #18 on github.

Author: Araneidae

cothread/py23.py

@@ -50,7 +50,7 @@ def auto_decode(result, func, args):
         if result is None:
             return result
         else:
-            return result.decode('UTF-8')
+            return result.decode('UTF-8', 'replace')
 
     def decode(s):
-        return s.decode('UTF-8')
+        return s.decode('UTF-8', 'replace')