Merge pull request #628 from fxdgear/nick/fix_encode_error

better handling of bytestrings in _escape

elasticsearch/client/utils.py

@@ -26,13 +26,13 @@ def _escape(value):
     elif isinstance(value, bool):
         value = str(value).lower()
 
+    # don't decode bytestrings
+    elif isinstance(value, bytes):
+        return value
+
     # encode strings to utf-8
     if isinstance(value, string_types):
-        try:
-            return value.encode('utf-8')
-        except UnicodeDecodeError:
-            # Python 2 and str, no need to re-encode
-            pass
+        return value.encode('utf-8')
 
     return str(value)
 

test_elasticsearch/test_client/test_utils.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
 from __future__ import unicode_literals
 
-from elasticsearch.client.utils import _make_path
+from elasticsearch.client.utils import _make_path, _escape
 from elasticsearch.compat import PY2
 
 from ..test_cases import TestCase, SkipTest
@@ -17,3 +17,24 @@ class TestMakePath(TestCase):
         id = "中文".encode('utf-8')
         self.assertEquals('/some-index/type/%E4%B8%AD%E6%96%87', _make_path('some-index', 'type', id))
 
+
+class TestEscape(TestCase):
+    def test_handles_ascii(self):
+        string = "abc123"
+        self.assertEquals(
+            b'abc123',
+            _escape(string)
+        )
+    def test_handles_unicode(self):
+        string = "中文"
+        self.assertEquals(
+            b'\xe4\xb8\xad\xe6\x96\x87',
+            _escape(string)
+        )
+
+    def test_handles_bytestring(self):
+        string = b'celery-task-meta-c4f1201f-eb7b-41d5-9318-a75a8cfbdaa0'
+        self.assertEquals(
+            string,
+            _escape(string)
+        )