commit 64b8f6ca15fcb88145c447562258ebfab992dd10
author Nobuaki Sukegawa <nsukeg@gmail.com> Sat Oct 10 02:12:48 2015 +0900
committer Roger Meier <roger@apache.org> Sun Oct 11 00:23:31 2015 +0200
tree 60e6c288699b265482a7b48095cc3a48540a3efd
parent 8cd519f7a3b9436ae95049ee4299441306bdeb5c

THRIFT-3375 Python TJSONProtocol encodes utf8 string values in an incompatible way

This closes #642

lib/py/src/protocol/TJSONProtocol.py

diff --git a/lib/py/src/protocol/TJSONProtocol.py b/lib/py/src/protocol/TJSONProtocol.py
index d270ce8..9e25f67 100644
--- a/lib/py/src/protocol/TJSONProtocol.py
+++ b/lib/py/src/protocol/TJSONProtocol.py
@@ -174,13 +174,13 @@
 
   def writeJSONString(self, string):
     self.context.write()
-    self.trans.write(json.dumps(string))
+    self.trans.write(json.dumps(string, ensure_ascii=False))
 
   def writeJSONNumber(self, number):
     self.context.write()
     jsNumber = str(number)
     if self.context.escapeNum():
-      jsNumber = "%s%s%s" % (QUOTE, jsNumber,  QUOTE)
+      jsNumber = "%s%s%s" % (QUOTE, jsNumber, QUOTE)
     self.trans.write(jsNumber)
 
   def writeJSONBase64(self, binary):

test/py/TestClient.py

diff --git a/test/py/TestClient.py b/test/py/TestClient.py
index 7fe48f6..7e3daf2 100755
--- a/test/py/TestClient.py
+++ b/test/py/TestClient.py
@@ -1,5 +1,5 @@
 #!/usr/bin/env python
-
+# -*- coding: utf-8 -*-
 #
 # Licensed to the Apache Software Foundation (ASF) under one
 # or more contributor license agreements. See the NOTICE file
@@ -105,6 +105,33 @@
     print('testString')
     self.assertEqual(self.client.testString('Python' * 20), 'Python' * 20)
     self.assertEqual(self.client.testString(''), '')
+    self.assertEqual(self.client.testString(u'パイソン'.encode('utf8')), u'パイソン'.encode('utf8'))
+    s = u"""Afrikaans, Alemannisch, Aragonés, العربية, مصرى,
+        Asturianu, Aymar aru, Azərbaycan, Башҡорт, Boarisch, Žemaitėška,
+        Беларуская, Беларуская (тарашкевіца), Български, Bamanankan,
+        বাংলা, Brezhoneg, Bosanski, Català, Mìng-dĕ̤ng-ngṳ̄, Нохчийн,
+        Cebuano, ᏣᎳᎩ, Česky, Словѣ́ньскъ / ⰔⰎⰑⰂⰡⰐⰠⰔⰍⰟ, Чӑвашла, Cymraeg,
+        Dansk, Zazaki, ދިވެހިބަސް, Ελληνικά, Emiliàn e rumagnòl, English,
+        Esperanto, Español, Eesti, Euskara, فارسی, Suomi, Võro, Føroyskt,
+        Français, Arpetan, Furlan, Frysk, Gaeilge, 贛語, Gàidhlig, Galego,
+        Avañe'ẽ, ગુજરાતી, Gaelg, עברית, हिन्दी, Fiji Hindi, Hrvatski,
+        Kreyòl ayisyen, Magyar, Հայերեն, Interlingua, Bahasa Indonesia,
+        Ilokano, Ido, Íslenska, Italiano, 日本語, Lojban, Basa Jawa,
+        ქართული, Kongo, Kalaallisut, ಕನ್ನಡ, 한국어, Къарачай-Малкъар,
+        Ripoarisch, Kurdî, Коми, Kernewek, Кыргызча, Latina, Ladino,
+        Lëtzebuergesch, Limburgs, Lingála, ລາວ, Lietuvių, Latviešu, Basa
+        Banyumasan, Malagasy, Македонски, മലയാളം, मराठी, مازِرونی, Bahasa
+        Melayu, Nnapulitano, Nedersaksisch, नेपाल भाषा, Nederlands, ‪
+        Norsk (nynorsk)‬, ‪Norsk (bokmål)‬, Nouormand, Diné bizaad,
+        Occitan, Иронау, Papiamentu, Deitsch, Polski, پنجابی, پښتو,
+        Norfuk / Pitkern, Português, Runa Simi, Rumantsch, Romani, Română,
+        Русский, Саха тыла, Sardu, Sicilianu, Scots, Sámegiella, Simple
+        English, Slovenčina, Slovenščina, Српски / Srpski, Seeltersk,
+        Svenska, Kiswahili, தமிழ், తెలుగు, Тоҷикӣ, ไทย, Türkmençe, Tagalog,
+        Türkçe, Татарча/Tatarça, Українська, اردو, Tiếng Việt, Volapük,
+        Walon, Winaray, 吴语, isiXhosa, ייִדיש, Yorùbá, Zeêuws, 中文,
+        Bân-lâm-gú, 粵語"""
+    self.assertEqual(self.client.testString(s.encode('utf8')), s.encode('utf8'))
 
   def testBool(self):
     print('testBool')