Fix unicode body handling.

Only body should be encoded as bytes, not anything going through the
serializer (mostly because of bulk)

elasticsearch/serializer.py

@@ -21,11 +21,7 @@ class JSONSerializer(object):
     def dumps(self, data):
         # don't serialize strings
         if isinstance(data, (type(''), type(u''))):
-            try:
-                return data.encode('utf-8')
-            except UnicodeDecodeError:
-                # Python 2 and str, no need to re-encode
-                return data
+            return data
 
         try:
             return json.dumps(data, default=self.default)

elasticsearch/transport.py

@@ -234,6 +234,13 @@ class Transport(object):
                     params['source'] = body
                     body = None
 
+        if body is not None:
+            try:
+                body = body.encode('utf-8')
+            except UnicodeDecodeError:
+                # Python 2 and str, no need to re-encode
+                pass
+
         ignore = ()
         if params and 'ignore' in params:
             ignore = params.pop('ignore')

test_elasticsearch/test_serializer.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 import sys
 
 from datetime import datetime
@@ -26,4 +27,4 @@ class TestJSONSerializer(TestCase):
         self.assertRaises(SerializationError, JSONSerializer().loads, '{{')
 
     def test_strings_are_left_untouched(self):
-        self.assertEquals('Hello World!', JSONSerializer().dumps('Hello World!'))
+        self.assertEquals(u"你好", JSONSerializer().dumps(u"你好"))

test_elasticsearch/test_transport.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 import time
 
 from elasticsearch.transport import Transport
@@ -47,7 +48,15 @@ class TestTransport(TestCase):
 
         t.perform_request('GET', '/', body={})
         self.assertEquals(1, len(t.get_connection().calls))
-        self.assertEquals(('POST', '/', None, '{}'), t.get_connection().calls[0][0])
+        self.assertEquals(('POST', '/', None, b'{}'), t.get_connection().calls[0][0])
+
+    def test_body_gets_encoded_into_bytes(self):
+        t = Transport([{}], connection_class=DummyConnection)
+
+        t.perform_request('GET', '/', body=u'你好')
+        self.assertEquals(1, len(t.get_connection().calls))
+        self.assertEquals(('GET', '/', None, b'\xe4\xbd\xa0\xe5\xa5\xbd'), t.get_connection().calls[0][0])
+
 
     def test_kwargs_passed_on_to_connections(self):
         t = Transport([{'host': 'google.com'}], port=123)