Fixed Python by updating failure lists and fixed a few broken tests.

Python 2.x doesn't detect unpaired surrogates so we have to
do that manually.

1 files changed, 10 insertions, 0 deletions

diff --git a/python/google/protobuf/json_format.py b/python/google/protobuf/json_format.py
index 57aa4077..8af6cd20 100644
--- a/python/google/protobuf/json_format.py
+++ b/python/google/protobuf/json_format.py
@@ -49,6 +49,7 @@ except ImportError:
 import base64
 import json
 import math
+import re
 import six
 import sys
 
@@ -68,6 +69,10 @@ _INFINITY = 'Infinity'
 _NEG_INFINITY = '-Infinity'
 _NAN = 'NaN'
 
+if sys.version_info < (3, 0):
+  _UNPAIRED_SURROGATE_PATTERN = re.compile(six.u(
+      r'[\ud800-\udbff](?![\udc00-\udfff])|(?<![\ud800-\udbff])[\udc00-\udfff]'
+  ))
 
 class Error(Exception):
   """Top-level module error for json_format."""
@@ -554,6 +559,11 @@ def _ConvertScalarFieldValue(value, field, require_str=False):
   elif field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_STRING:
     if field.type == descriptor.FieldDescriptor.TYPE_BYTES:
       return base64.b64decode(value)
+    elif sys.version_info < (3, 0):
+      # Python 2.x does not detect unpaired surrogates when JSON parsing.
+      if _UNPAIRED_SURROGATE_PATTERN.search(value):
+        raise ParseError('Unpaired surrogate')
+      return value
     else:
       return value
   elif field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_ENUM: