Merge remote-tracking branch 'origin/master' into beta-1

.gitignore

@@ -55,6 +55,7 @@ src/google/protobuf/util/**/*.pb.h
 *_pb2.py
 python/*.egg
 python/.eggs/
+python/.tox
 python/build/
 python/google/protobuf/compiler/
 

Makefile.am

@@ -538,7 +538,6 @@ python_EXTRA_DIST=                                                           \
   python/google/protobuf/text_format.py                                      \
   python/google/protobuf/__init__.py                                         \
   python/google/__init__.py                                                  \
-  python/ez_setup.py                                                         \
   python/setup.py                                                            \
   python/mox.py                                                              \
   python/stubout.py                                                          \

csharp/generate_protos.sh

@@ -36,11 +36,8 @@ if [ -z "$PROTOC" ]; then
 fi
 
 # Descriptor proto
-# TODO(jonskeet): Remove fixup
-cp src/google/protobuf/descriptor.proto src/google/protobuf/descriptor_proto_file.proto
 $PROTOC -Isrc --csharp_out=csharp/src/Google.Protobuf/Reflection \
-    src/google/protobuf/descriptor_proto_file.proto
-rm src/google/protobuf/descriptor_proto_file.proto
+    src/google/protobuf/descriptor.proto
 
 $PROTOC -Isrc --csharp_out=csharp/src/Google.Protobuf/WellKnownTypes \
     src/google/protobuf/any.proto \

csharp/src/Google.Protobuf/Reflection/FileDescriptor.cs

@@ -300,8 +300,9 @@ namespace Google.Protobuf.Reflection
                 if (dependencies[i].Name != proto.Dependency[i])
                 {
                     throw new DescriptorValidationException(result,
-                                                            "Dependencies passed to FileDescriptor.BuildFrom() don't match " +
-                                                            "those listed in the FileDescriptorProto.");
+                        "Dependencies passed to FileDescriptor.BuildFrom() don't match " +
+                        "those listed in the FileDescriptorProto. Expected: " +
+                        proto.Dependency[i] + " but was: " + dependencies[i].Name);
                 }
             }
 

python/google/protobuf/__init__.py

@@ -28,8 +28,6 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-# Needs to stay compatible with Python 2.5 due to GAE.
-#
 # Copyright 2007 Google Inc. All Rights Reserved.
 
 __version__ = '3.0.0a4.dev0'

python/google/protobuf/descriptor.py

@@ -28,8 +28,6 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-# Needs to stay compatible with Python 2.5 due to GAE.
-#
 # Copyright 2007 Google Inc. All Rights Reserved.
 
 """Descriptors essentially contain exactly the information found in a .proto
@@ -918,5 +916,5 @@ def MakeDescriptor(desc_proto, package='', build_file_if_cpp=True,
 
   desc_name = '.'.join(full_message_name)
   return Descriptor(desc_proto.name, desc_name, None, None, fields,
-                    nested_types.values(), enum_types.values(), [],
+                    list(nested_types.values()), list(enum_types.values()), [],
                     options=desc_proto.options)

python/google/protobuf/descriptor_pool.py

@@ -57,8 +57,6 @@ directly instead of this class.
 
 __author__ = 'matthewtoia@google.com (Matt Toia)'
 
-import sys
-
 from google.protobuf import descriptor
 from google.protobuf import descriptor_database
 from google.protobuf import text_encoding
@@ -192,8 +190,7 @@ class DescriptorPool(object):
 
     try:
       file_proto = self._internal_db.FindFileByName(file_name)
-    except KeyError:
-      _, error, _ = sys.exc_info()  #PY25 compatible for GAE.
+    except KeyError as error:
       if self._descriptor_db:
         file_proto = self._descriptor_db.FindFileByName(file_name)
       else:
@@ -228,8 +225,7 @@ class DescriptorPool(object):
 
     try:
       file_proto = self._internal_db.FindFileContainingSymbol(symbol)
-    except KeyError:
-      _, error, _ = sys.exc_info()  #PY25 compatible for GAE.
+    except KeyError as error:
       if self._descriptor_db:
         file_proto = self._descriptor_db.FindFileContainingSymbol(symbol)
       else:

python/google/protobuf/internal/_parameterized.py

@@ -43,7 +43,7 @@ A simple example:
        (4, 5, 9),
        (1, 1, 3))
     def testAddition(self, op1, op2, result):
-      self.assertEquals(result, op1 + op2)
+      self.assertEqual(result, op1 + op2)
 
 
 Each invocation is a separate test case and properly isolated just
@@ -60,7 +60,7 @@ or dictionaries (with named parameters):
        {'op1': 4, 'op2': 5, 'result': 9},
     )
     def testAddition(self, op1, op2, result):
-      self.assertEquals(result, op1 + op2)
+      self.assertEqual(result, op1 + op2)
 
 If a parameterized test fails, the error message will show the
 original test name (which is modified internally) and the arguments
@@ -88,7 +88,7 @@ str()):
        ('EmptyPrefix', '', 'abc', True),
        ('BothEmpty', '', '', True))
     def testStartsWith(self, prefix, string, result):
-      self.assertEquals(result, strings.startswith(prefix))
+      self.assertEqual(result, strings.startswith(prefix))
 
 Named tests also have the benefit that they can be run individually
 from the command line:
@@ -127,7 +127,7 @@ the decorator. This iterable will be used to obtain the test cases:
       c.op1, c.op2, c.result for c in testcases
     )
     def testAddition(self, op1, op2, result):
-      self.assertEquals(result, op1 + op2)
+      self.assertEqual(result, op1 + op2)
 
 
 Single-Argument Test Methods
@@ -149,9 +149,14 @@ import collections
 import functools
 import re
 import types
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 import uuid
 
+import six
+
 ADDR_RE = re.compile(r'\<([a-zA-Z0-9_\-\.]+) object at 0x[a-fA-F0-9]+\>')
 _SEPARATOR = uuid.uuid1().hex
 _FIRST_ARG = object()
@@ -170,13 +175,13 @@ def _StrClass(cls):
 
 def _NonStringIterable(obj):
   return (isinstance(obj, collections.Iterable) and not
-          isinstance(obj, basestring))
+          isinstance(obj, six.string_types))
 
 
 def _FormatParameterList(testcase_params):
   if isinstance(testcase_params, collections.Mapping):
     return ', '.join('%s=%s' % (argname, _CleanRepr(value))
-                     for argname, value in testcase_params.iteritems())
+                     for argname, value in testcase_params.items())
   elif _NonStringIterable(testcase_params):
     return ', '.join(map(_CleanRepr, testcase_params))
   else:
@@ -258,7 +263,9 @@ def _ModifyClass(class_object, testcases, naming_type):
       'Cannot add parameters to %s,'
       ' which already has parameterized methods.' % (class_object,))
   class_object._id_suffix = id_suffix = {}
-  for name, obj in class_object.__dict__.items():
+  # We change the size of __dict__ while we iterate over it, 
+  # which Python 3.x will complain about, so use copy().
+  for name, obj in class_object.__dict__.copy().items():
     if (name.startswith(unittest.TestLoader.testMethodPrefix)
         and isinstance(obj, types.FunctionType)):
       delattr(class_object, name)
@@ -266,7 +273,7 @@ def _ModifyClass(class_object, testcases, naming_type):
       _UpdateClassDictForParamTestCase(
           methods, id_suffix, name,
           _ParameterizedTestIter(obj, testcases, naming_type))
-      for name, meth in methods.iteritems():
+      for name, meth in methods.items():
         setattr(class_object, name, meth)
 
 

python/google/protobuf/internal/decoder.py

@@ -28,8 +28,6 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-#PY25 compatible for GAE.
-#
 # Copyright 2009 Google Inc. All Rights Reserved.
 
 """Code for decoding protocol buffer primitives.
@@ -85,8 +83,12 @@ we repeatedly read a tag, look up the corresponding decoder, and invoke it.
 __author__ = 'kenton@google.com (Kenton Varda)'
 
 import struct
-import sys  ##PY25
-_PY2 = sys.version_info[0] < 3  ##PY25
+
+import six
+
+if six.PY3:
+  long = int
+
 from google.protobuf.internal import encoder
 from google.protobuf.internal import wire_format
 from google.protobuf import message
@@ -114,14 +116,11 @@ def _VarintDecoder(mask, result_type):
   decoder returns a (value, new_pos) pair.
   """
 
-  local_ord = ord
-  py2 = _PY2  ##PY25
-##!PY25  py2 = str is bytes
   def DecodeVarint(buffer, pos):
     result = 0
     shift = 0
     while 1:
-      b = local_ord(buffer[pos]) if py2 else buffer[pos]
+      b = six.indexbytes(buffer, pos)
       result |= ((b & 0x7f) << shift)
       pos += 1
       if not (b & 0x80):
@@ -137,14 +136,11 @@ def _VarintDecoder(mask, result_type):
 def _SignedVarintDecoder(mask, result_type):
   """Like _VarintDecoder() but decodes signed values."""
 
-  local_ord = ord
-  py2 = _PY2  ##PY25
-##!PY25  py2 = str is bytes
   def DecodeVarint(buffer, pos):
     result = 0
     shift = 0
     while 1:
-      b = local_ord(buffer[pos]) if py2 else buffer[pos]
+      b = six.indexbytes(buffer, pos)
       result |= ((b & 0x7f) << shift)
       pos += 1
       if not (b & 0x80):
@@ -183,10 +179,8 @@ def ReadTag(buffer, pos):
   use that, but not in Python.
   """
 
-  py2 = _PY2  ##PY25
-##!PY25  py2 = str is bytes
   start = pos
-  while (ord(buffer[pos]) if py2 else buffer[pos]) & 0x80:
+  while six.indexbytes(buffer, pos) & 0x80:
     pos += 1
   pos += 1
   return (buffer[start:pos], pos)
@@ -301,7 +295,6 @@ def _FloatDecoder():
   """
 
   local_unpack = struct.unpack
-  b = (lambda x:x) if _PY2 else lambda x:x.encode('latin1')  ##PY25
 
   def InnerDecode(buffer, pos):
     # We expect a 32-bit value in little-endian byte order.  Bit 1 is the sign
@@ -312,17 +305,12 @@ def _FloatDecoder():
     # If this value has all its exponent bits set, then it's non-finite.
     # In Python 2.4, struct.unpack will convert it to a finite 64-bit value.
     # To avoid that, we parse it specially.
-    if ((float_bytes[3:4] in b('\x7F\xFF'))  ##PY25
-##!PY25    if ((float_bytes[3:4] in b'\x7F\xFF')
-        and (float_bytes[2:3] >= b('\x80'))):  ##PY25
-##!PY25        and (float_bytes[2:3] >= b'\x80')):
+    if (float_bytes[3:4] in b'\x7F\xFF' and float_bytes[2:3] >= b'\x80'):
       # If at least one significand bit is set...
-      if float_bytes[0:3] != b('\x00\x00\x80'):  ##PY25
-##!PY25      if float_bytes[0:3] != b'\x00\x00\x80':
+      if float_bytes[0:3] != b'\x00\x00\x80':
         return (_NAN, new_pos)
       # If sign bit is set...
-      if float_bytes[3:4] == b('\xFF'):  ##PY25
-##!PY25      if float_bytes[3:4] == b'\xFF':
+      if float_bytes[3:4] == b'\xFF':
         return (_NEG_INF, new_pos)
       return (_POS_INF, new_pos)
 
@@ -341,7 +329,6 @@ def _DoubleDecoder():
   """
 
   local_unpack = struct.unpack
-  b = (lambda x:x) if _PY2 else lambda x:x.encode('latin1')  ##PY25
 
   def InnerDecode(buffer, pos):
     # We expect a 64-bit value in little-endian byte order.  Bit 1 is the sign
@@ -352,12 +339,9 @@ def _DoubleDecoder():
     # If this value has all its exponent bits set and at least one significand
     # bit set, it's not a number.  In Python 2.4, struct.unpack will treat it
     # as inf or -inf.  To avoid that, we treat it specially.
-##!PY25    if ((double_bytes[7:8] in b'\x7F\xFF')
-##!PY25        and (double_bytes[6:7] >= b'\xF0')
-##!PY25        and (double_bytes[0:7] != b'\x00\x00\x00\x00\x00\x00\xF0')):
-    if ((double_bytes[7:8] in b('\x7F\xFF'))  ##PY25
-        and (double_bytes[6:7] >= b('\xF0'))  ##PY25
-        and (double_bytes[0:7] != b('\x00\x00\x00\x00\x00\x00\xF0'))):  ##PY25
+    if ((double_bytes[7:8] in b'\x7F\xFF')
+        and (double_bytes[6:7] >= b'\xF0')
+        and (double_bytes[0:7] != b'\x00\x00\x00\x00\x00\x00\xF0')):
       return (_NAN, new_pos)
 
     # Note that we expect someone up-stack to catch struct.error and convert
@@ -480,12 +464,12 @@ def StringDecoder(field_number, is_repeated, is_packed, key, new_default):
   """Returns a decoder for a string field."""
 
   local_DecodeVarint = _DecodeVarint
-  local_unicode = unicode
+  local_unicode = six.text_type
 
   def _ConvertToUnicode(byte_str):
     try:
       return local_unicode(byte_str, 'utf-8')
-    except UnicodeDecodeError, e:
+    except UnicodeDecodeError as e:
       # add more information to the error message and re-raise it.
       e.reason = '%s in field: %s' % (e, key.full_name)
       raise

python/google/protobuf/internal/descriptor_database_test.py

@@ -34,7 +34,10 @@
 
 __author__ = 'matthewtoia@google.com (Matt Toia)'
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf import descriptor_pb2
 from google.protobuf.internal import factory_test2_pb2
 from google.protobuf import descriptor_database
@@ -48,17 +51,17 @@ class DescriptorDatabaseTest(unittest.TestCase):
         factory_test2_pb2.DESCRIPTOR.serialized_pb)
     db.Add(file_desc_proto)
 
-    self.assertEquals(file_desc_proto, db.FindFileByName(
+    self.assertEqual(file_desc_proto, db.FindFileByName(
         'google/protobuf/internal/factory_test2.proto'))
-    self.assertEquals(file_desc_proto, db.FindFileContainingSymbol(
+    self.assertEqual(file_desc_proto, db.FindFileContainingSymbol(
         'google.protobuf.python.internal.Factory2Message'))
-    self.assertEquals(file_desc_proto, db.FindFileContainingSymbol(
+    self.assertEqual(file_desc_proto, db.FindFileContainingSymbol(
         'google.protobuf.python.internal.Factory2Message.NestedFactory2Message'))
-    self.assertEquals(file_desc_proto, db.FindFileContainingSymbol(
+    self.assertEqual(file_desc_proto, db.FindFileContainingSymbol(
         'google.protobuf.python.internal.Factory2Enum'))
-    self.assertEquals(file_desc_proto, db.FindFileContainingSymbol(
+    self.assertEqual(file_desc_proto, db.FindFileContainingSymbol(
         'google.protobuf.python.internal.Factory2Message.NestedFactory2Enum'))
-    self.assertEquals(file_desc_proto, db.FindFileContainingSymbol(
+    self.assertEqual(file_desc_proto, db.FindFileContainingSymbol(
         'google.protobuf.python.internal.MessageWithNestedEnumOnly.NestedEnum'))
 
 if __name__ == '__main__':

python/google/protobuf/internal/descriptor_test.py

-#! /usr/bin/env python
+#! /usr/bin/python
 #
 # Protocol Buffers - Google's data interchange format
 # Copyright 2008 Google Inc.  All rights reserved.
@@ -36,7 +36,6 @@ __author__ = 'robinson@google.com (Will Robinson)'
 
 import sys
 
-import unittest
 from google.protobuf import unittest_custom_options_pb2
 from google.protobuf import unittest_import_pb2
 from google.protobuf import unittest_pb2
@@ -46,6 +45,11 @@ from google.protobuf import descriptor
 from google.protobuf import symbol_database
 from google.protobuf import text_format
 
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
+
 
 TEST_EMPTY_MESSAGE_DESCRIPTOR_ASCII = """
 name: 'TestEmptyMessage'
@@ -455,7 +459,7 @@ class GeneratedDescriptorTest(unittest.TestCase):
     # properties of an immutable abc.Mapping.
     self.assertGreater(len(mapping), 0)  # Sized
     self.assertEqual(len(mapping), len(list(mapping)))  # Iterable
-    if sys.version_info.major >= 3:
+    if sys.version_info >= (3,):
       key, item = next(iter(mapping.items()))
     else:
       key, item = mapping.items()[0]
@@ -464,7 +468,7 @@ class GeneratedDescriptorTest(unittest.TestCase):
     # keys(), iterkeys() &co
     item = (next(iter(mapping.keys())), next(iter(mapping.values())))
     self.assertEqual(item, next(iter(mapping.items())))
-    if sys.version_info.major < 3:
+    if sys.version_info < (3,):
       def CheckItems(seq, iterator):
         self.assertEqual(next(iterator), seq[0])
         self.assertEqual(list(iterator), seq[1:])
@@ -772,7 +776,7 @@ class MakeDescriptorTest(unittest.TestCase):
     reformed_descriptor = descriptor.MakeDescriptor(descriptor_proto)
 
     options = reformed_descriptor.GetOptions()
-    self.assertEquals(101,
+    self.assertEqual(101,
                       options.Extensions[unittest_custom_options_pb2.msgopt].i)
 
 if __name__ == '__main__':

python/google/protobuf/internal/encoder.py

@@ -28,8 +28,6 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-#PY25 compatible for GAE.
-#
 # Copyright 2009 Google Inc. All Rights Reserved.
 
 """Code for encoding protocol message primitives.
@@ -45,7 +43,7 @@ FieldDescriptor) we construct two functions:  a "sizer" and an "encoder".  The
 sizer takes a value of this field's type and computes its byte size.  The
 encoder takes a writer function and a value.  It encodes the value into byte
 strings and invokes the writer function to write those strings.  Typically the
-writer function is the write() method of a cStringIO.
+writer function is the write() method of a BytesIO.
 
 We try to do as much work as possible when constructing the writer and the
 sizer rather than when calling them.  In particular:
@@ -71,8 +69,9 @@ sizer rather than when calling them.  In particular:
 __author__ = 'kenton@google.com (Kenton Varda)'
 
 import struct
-import sys  ##PY25
-_PY2 = sys.version_info[0] < 3  ##PY25
+
+import six
+
 from google.protobuf.internal import wire_format
 
 
@@ -372,16 +371,14 @@ def MapSizer(field_descriptor):
 def _VarintEncoder():
   """Return an encoder for a basic varint value (does not include tag)."""
 
-  local_chr = _PY2 and chr or (lambda x: bytes((x,)))  ##PY25
-##!PY25  local_chr = chr if bytes is str else lambda x: bytes((x,))
   def EncodeVarint(write, value):
     bits = value & 0x7f
     value >>= 7
     while value:
-      write(local_chr(0x80|bits))
+      write(six.int2byte(0x80|bits))
       bits = value & 0x7f
       value >>= 7
-    return write(local_chr(bits))
+    return write(six.int2byte(bits))
 
   return EncodeVarint
 
@@ -390,18 +387,16 @@ def _SignedVarintEncoder():
   """Return an encoder for a basic signed varint value (does not include
   tag)."""
 
-  local_chr = _PY2 and chr or (lambda x: bytes((x,)))  ##PY25
-##!PY25  local_chr = chr if bytes is str else lambda x: bytes((x,))
   def EncodeSignedVarint(write, value):
     if value < 0:
       value += (1 << 64)
     bits = value & 0x7f
     value >>= 7
     while value:
-      write(local_chr(0x80|bits))
+      write(six.int2byte(0x80|bits))
       bits = value & 0x7f
       value >>= 7
-    return write(local_chr(bits))
+    return write(six.int2byte(bits))
 
   return EncodeSignedVarint
 
@@ -416,8 +411,7 @@ def _VarintBytes(value):
 
   pieces = []
   _EncodeVarint(pieces.append, value)
-  return "".encode("latin1").join(pieces)  ##PY25
-##!PY25  return b"".join(pieces)
+  return b"".join(pieces)
 
 
 def TagBytes(field_number, wire_type):
@@ -555,33 +549,26 @@ def _FloatingPointEncoder(wire_type, format):
       format:  The format string to pass to struct.pack().
   """
 
-  b = _PY2 and (lambda x:x) or (lambda x:x.encode('latin1'))  ##PY25
   value_size = struct.calcsize(format)
   if value_size == 4:
     def EncodeNonFiniteOrRaise(write, value):
       # Remember that the serialized form uses little-endian byte order.
       if value == _POS_INF:
-        write(b('\x00\x00\x80\x7F'))  ##PY25
-##!PY25        write(b'\x00\x00\x80\x7F')
+        write(b'\x00\x00\x80\x7F')
       elif value == _NEG_INF:
-        write(b('\x00\x00\x80\xFF'))  ##PY25
-##!PY25        write(b'\x00\x00\x80\xFF')
+        write(b'\x00\x00\x80\xFF')
       elif value != value:           # NaN
-        write(b('\x00\x00\xC0\x7F'))  ##PY25
-##!PY25        write(b'\x00\x00\xC0\x7F')
+        write(b'\x00\x00\xC0\x7F')
       else:
         raise
   elif value_size == 8:
     def EncodeNonFiniteOrRaise(write, value):
       if value == _POS_INF:
-        write(b('\x00\x00\x00\x00\x00\x00\xF0\x7F'))  ##PY25
-##!PY25        write(b'\x00\x00\x00\x00\x00\x00\xF0\x7F')
+        write(b'\x00\x00\x00\x00\x00\x00\xF0\x7F')
       elif value == _NEG_INF:
-        write(b('\x00\x00\x00\x00\x00\x00\xF0\xFF'))  ##PY25
-##!PY25        write(b'\x00\x00\x00\x00\x00\x00\xF0\xFF')
+        write(b'\x00\x00\x00\x00\x00\x00\xF0\xFF')
       elif value != value:                         # NaN
-        write(b('\x00\x00\x00\x00\x00\x00\xF8\x7F'))  ##PY25
-##!PY25        write(b'\x00\x00\x00\x00\x00\x00\xF8\x7F')
+        write(b'\x00\x00\x00\x00\x00\x00\xF8\x7F')
       else:
         raise
   else:
@@ -657,10 +644,8 @@ DoubleEncoder   = _FloatingPointEncoder(wire_format.WIRETYPE_FIXED64, '<d')
 def BoolEncoder(field_number, is_repeated, is_packed):
   """Returns an encoder for a boolean field."""
 
-##!PY25  false_byte = b'\x00'
-##!PY25  true_byte = b'\x01'
-  false_byte = '\x00'.encode('latin1')  ##PY25
-  true_byte = '\x01'.encode('latin1')  ##PY25
+  false_byte = b'\x00'
+  true_byte = b'\x01'
   if is_packed:
     tag_bytes = TagBytes(field_number, wire_format.WIRETYPE_LENGTH_DELIMITED)
     local_EncodeVarint = _EncodeVarint
@@ -796,8 +781,7 @@ def MessageSetItemEncoder(field_number):
       }
     }
   """
-  start_bytes = "".encode("latin1").join([  ##PY25
-##!PY25  start_bytes = b"".join([
+  start_bytes = b"".join([
       TagBytes(1, wire_format.WIRETYPE_START_GROUP),
       TagBytes(2, wire_format.WIRETYPE_VARINT),
       _VarintBytes(field_number),

python/google/protobuf/internal/generator_test.py

@@ -41,7 +41,10 @@ further ensures that we can use Python protocol message objects as we expect.
 
 __author__ = 'robinson@google.com (Will Robinson)'
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf.internal import test_bad_identifiers_pb2
 from google.protobuf import unittest_custom_options_pb2
 from google.protobuf import unittest_import_pb2
@@ -154,7 +157,7 @@ class GeneratorTest(unittest.TestCase):
     # extension and for its value to be set to -789.
 
   def testNestedTypes(self):
-    self.assertEquals(
+    self.assertEqual(
         set(unittest_pb2.TestAllTypes.DESCRIPTOR.nested_types),
         set([
             unittest_pb2.TestAllTypes.NestedMessage.DESCRIPTOR,
@@ -292,10 +295,10 @@ class GeneratorTest(unittest.TestCase):
     self.assertIs(desc.oneofs[0], desc.oneofs_by_name['oneof_field'])
     nested_names = set(['oneof_uint32', 'oneof_nested_message',
                         'oneof_string', 'oneof_bytes'])
-    self.assertItemsEqual(
+    self.assertEqual(
         nested_names,
-        [field.name for field in desc.oneofs[0].fields])
-    for field_name, field_desc in desc.fields_by_name.iteritems():
+        set([field.name for field in desc.oneofs[0].fields]))
+    for field_name, field_desc in desc.fields_by_name.items():
       if field_name in nested_names:
         self.assertIs(desc.oneofs[0], field_desc.containing_oneof)
       else:
@@ -306,36 +309,36 @@ class SymbolDatabaseRegistrationTest(unittest.TestCase):
   """Checks that messages, enums and files are correctly registered."""
 
   def testGetSymbol(self):
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes, symbol_database.Default().GetSymbol(
             'protobuf_unittest.TestAllTypes'))
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes.NestedMessage,
         symbol_database.Default().GetSymbol(
             'protobuf_unittest.TestAllTypes.NestedMessage'))
     with self.assertRaises(KeyError):
       symbol_database.Default().GetSymbol('protobuf_unittest.NestedMessage')
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes.OptionalGroup,
         symbol_database.Default().GetSymbol(
             'protobuf_unittest.TestAllTypes.OptionalGroup'))
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes.RepeatedGroup,
         symbol_database.Default().GetSymbol(
             'protobuf_unittest.TestAllTypes.RepeatedGroup'))
 
   def testEnums(self):
-    self.assertEquals(
+    self.assertEqual(
         'protobuf_unittest.ForeignEnum',
         symbol_database.Default().pool.FindEnumTypeByName(
             'protobuf_unittest.ForeignEnum').full_name)
-    self.assertEquals(
+    self.assertEqual(
         'protobuf_unittest.TestAllTypes.NestedEnum',
         symbol_database.Default().pool.FindEnumTypeByName(
             'protobuf_unittest.TestAllTypes.NestedEnum').full_name)
 
   def testFindFileByName(self):
-    self.assertEquals(
+    self.assertEqual(
         'google/protobuf/unittest.proto',
         symbol_database.Default().pool.FindFileByName(
             'google/protobuf/unittest.proto').name)

python/google/protobuf/internal/message_factory_test.py

@@ -34,7 +34,10 @@
 
 __author__ = 'matthewtoia@google.com (Matt Toia)'
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf import descriptor_pb2
 from google.protobuf.internal import factory_test1_pb2
 from google.protobuf.internal import factory_test2_pb2
@@ -81,9 +84,9 @@ class MessageFactoryTest(unittest.TestCase):
     serialized = msg.SerializeToString()
     converted = factory_test2_pb2.Factory2Message.FromString(serialized)
     reserialized = converted.SerializeToString()
-    self.assertEquals(serialized, reserialized)
+    self.assertEqual(serialized, reserialized)
     result = cls.FromString(reserialized)
-    self.assertEquals(msg, result)
+    self.assertEqual(msg, result)
 
   def testGetPrototype(self):
     db = descriptor_database.DescriptorDatabase()
@@ -93,11 +96,11 @@ class MessageFactoryTest(unittest.TestCase):
     factory = message_factory.MessageFactory()
     cls = factory.GetPrototype(pool.FindMessageTypeByName(
         'google.protobuf.python.internal.Factory2Message'))
-    self.assertIsNot(cls, factory_test2_pb2.Factory2Message)
+    self.assertFalse(cls is factory_test2_pb2.Factory2Message)
     self._ExerciseDynamicClass(cls)
     cls2 = factory.GetPrototype(pool.FindMessageTypeByName(
         'google.protobuf.python.internal.Factory2Message'))
-    self.assertIs(cls, cls2)
+    self.assertTrue(cls is cls2)
 
   def testGetMessages(self):
     # performed twice because multiple calls with the same input must be allowed
@@ -124,8 +127,8 @@ class MessageFactoryTest(unittest.TestCase):
           'google.protobuf.python.internal.another_field']
       msg1.Extensions[ext1] = 'test1'
       msg1.Extensions[ext2] = 'test2'
-      self.assertEquals('test1', msg1.Extensions[ext1])
-      self.assertEquals('test2', msg1.Extensions[ext2])
+      self.assertEqual('test1', msg1.Extensions[ext1])
+      self.assertEqual('test2', msg1.Extensions[ext2])
 
 
 if __name__ == '__main__':

python/google/protobuf/internal/message_test.py

@@ -49,9 +49,16 @@ import math
 import operator
 import pickle
 import sys
-import unittest
 
-import unittest
+import six
+
+if six.PY3:
+  long = int
+
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf.internal import _parameterized
 from google.protobuf import map_unittest_pb2
 from google.protobuf import unittest_pb2
@@ -322,7 +329,7 @@ class MessageTest(unittest.TestCase):
   def testHighPrecisionFloatPrinting(self, message_module):
     message = message_module.TestAllTypes()
     message.optional_double = 0.12345678912345678
-    if sys.version_info.major >= 3:
+    if sys.version_info >= (3,):
       self.assertEqual(str(message), 'optional_double: 0.12345678912345678\n')
     else:
       self.assertEqual(str(message), 'optional_double: 0.123456789123\n')
@@ -466,7 +473,7 @@ class MessageTest(unittest.TestCase):
     message.repeated_nested_message.sort(key=get_bb, reverse=True)
     self.assertEqual([k.bb for k in message.repeated_nested_message],
                      [6, 5, 4, 3, 2, 1])
-    if sys.version_info.major >= 3: return  # No cmp sorting in PY3.
+    if sys.version_info >= (3,): return  # No cmp sorting in PY3.
     message.repeated_nested_message.sort(sort_function=cmp_bb)
     self.assertEqual([k.bb for k in message.repeated_nested_message],
                      [1, 2, 3, 4, 5, 6])
@@ -485,7 +492,7 @@ class MessageTest(unittest.TestCase):
     self.assertEqual(list(message.repeated_int32), [-1, -2, -3])
     message.repeated_int32.sort(key=abs, reverse=True)
     self.assertEqual(list(message.repeated_int32), [-3, -2, -1])
-    if sys.version_info.major < 3:  # No cmp sorting in PY3.
+    if sys.version_info < (3,):  # No cmp sorting in PY3.
       abs_cmp = lambda a, b: cmp(abs(a), abs(b))
       message.repeated_int32.sort(sort_function=abs_cmp)
       self.assertEqual(list(message.repeated_int32), [-1, -2, -3])
@@ -499,7 +506,7 @@ class MessageTest(unittest.TestCase):
     self.assertEqual(list(message.repeated_string), ['c', 'bb', 'aaa'])
     message.repeated_string.sort(key=len, reverse=True)
     self.assertEqual(list(message.repeated_string), ['aaa', 'bb', 'c'])
-    if sys.version_info.major < 3:  # No cmp sorting in PY3.
+    if sys.version_info < (3,):  # No cmp sorting in PY3.
       len_cmp = lambda a, b: cmp(len(a), len(b))
       message.repeated_string.sort(sort_function=len_cmp)
       self.assertEqual(list(message.repeated_string), ['c', 'bb', 'aaa'])
@@ -522,7 +529,7 @@ class MessageTest(unittest.TestCase):
     m2.repeated_nested_message.add().bb = 2
     m2.repeated_nested_message.add().bb = 3
 
-    if sys.version_info.major >= 3: return  # No cmp() in PY3.
+    if sys.version_info >= (3,): return  # No cmp() in PY3.
 
     # These comparisons should not raise errors.
     _ = m1 < m2
@@ -721,9 +728,7 @@ class MessageTest(unittest.TestCase):
     in the value being converted to a Unicode string."""
     m = message_module.TestAllTypes()
     m.optional_string = str('')
-    self.assertTrue(isinstance(m.optional_string, unicode))
-
-# TODO(haberman): why are these tests Google-internal only?
+    self.assertIsInstance(m.optional_string, six.text_type)
 
   def testLongValuedSlice(self, message_module):
     """It should be possible to use long-valued indicies in slices
@@ -1109,14 +1114,13 @@ class Proto2Test(unittest.TestCase):
         repeated_nested_enum=['FOO', unittest_pb2.TestAllTypes.BAR],
         default_int32=800,
         oneof_string='y')
-    self.assertTrue(isinstance(message, unittest_pb2.TestAllTypes))
+    self.assertIsInstance(message, unittest_pb2.TestAllTypes)
     self.assertEqual(100, message.optional_int32)
     self.assertEqual(200, message.optional_fixed32)
     self.assertEqual(300.5, message.optional_float)
     self.assertEqual(b'x', message.optional_bytes)
     self.assertEqual(400, message.optionalgroup.a)
-    self.assertTrue(isinstance(message.optional_nested_message,
-                               unittest_pb2.TestAllTypes.NestedMessage))
+    self.assertIsInstance(message.optional_nested_message, unittest_pb2.TestAllTypes.NestedMessage)
     self.assertEqual(500, message.optional_nested_message.bb)
     self.assertEqual(unittest_pb2.TestAllTypes.BAZ,
                      message.optional_nested_enum)
@@ -1274,7 +1278,7 @@ class Proto3Test(unittest.TestCase):
     self.assertTrue('abc' in msg.map_string_string)
     self.assertTrue(888 in msg.map_int32_enum)
 
-    self.assertTrue(isinstance(msg.map_string_string['abc'], unicode))
+    self.assertIsInstance(msg.map_string_string['abc'], six.text_type)
 
     # Accessing an unset key still throws TypeError of the type of the key
     # is incorrect.
@@ -1289,14 +1293,14 @@ class Proto3Test(unittest.TestCase):
     msg = map_unittest_pb2.TestMap()
 
     self.assertIsNone(msg.map_int32_int32.get(5))
-    self.assertEquals(10, msg.map_int32_int32.get(5, 10))
+    self.assertEqual(10, msg.map_int32_int32.get(5, 10))
     self.assertIsNone(msg.map_int32_int32.get(5))
 
     msg.map_int32_int32[5] = 15
-    self.assertEquals(15, msg.map_int32_int32.get(5))
+    self.assertEqual(15, msg.map_int32_int32.get(5))
 
     self.assertIsNone(msg.map_int32_foreign_message.get(5))
-    self.assertEquals(10, msg.map_int32_foreign_message.get(5, 10))
+    self.assertEqual(10, msg.map_int32_foreign_message.get(5, 10))
 
     submsg = msg.map_int32_foreign_message[5]
     self.assertIs(submsg, msg.map_int32_foreign_message.get(5))
@@ -1353,17 +1357,17 @@ class Proto3Test(unittest.TestCase):
     msg = map_unittest_pb2.TestMap()
 
     unicode_obj = u'\u1234'
-    bytes_obj = unicode_obj.encode('utf8') 
+    bytes_obj = unicode_obj.encode('utf8')
 
     msg.map_string_string[bytes_obj] = bytes_obj
 
-    (key, value) = msg.map_string_string.items()[0]
+    (key, value) = list(msg.map_string_string.items())[0]
 
     self.assertEqual(key, unicode_obj)
     self.assertEqual(value, unicode_obj)
 
-    self.assertTrue(isinstance(key, unicode))
-    self.assertTrue(isinstance(value, unicode))
+    self.assertIsInstance(key, six.text_type)
+    self.assertIsInstance(value, six.text_type)
 
   def testMessageMap(self):
     msg = map_unittest_pb2.TestMap()
@@ -1531,7 +1535,7 @@ class Proto3Test(unittest.TestCase):
 
     submsg = msg.map_int32_foreign_message[111]
     self.assertIs(submsg, msg.map_int32_foreign_message[111])
-    self.assertTrue(isinstance(submsg, unittest_pb2.ForeignMessage))
+    self.assertIsInstance(submsg, unittest_pb2.ForeignMessage)
 
     submsg.c = 5
 
@@ -1548,7 +1552,7 @@ class Proto3Test(unittest.TestCase):
   def testMapIteration(self):
     msg = map_unittest_pb2.TestMap()
 
-    for k, v in msg.map_int32_int32.iteritems():
+    for k, v in msg.map_int32_int32.items():
       # Should not be reached.
       self.assertTrue(False)
 
@@ -1558,7 +1562,7 @@ class Proto3Test(unittest.TestCase):
     self.assertEqual(3, len(msg.map_int32_int32))
 
     matching_dict = {2: 4, 3: 6, 4: 8}
-    self.assertMapIterEquals(msg.map_int32_int32.iteritems(), matching_dict)
+    self.assertMapIterEquals(msg.map_int32_int32.items(), matching_dict)
 
   def testMapIterationClearMessage(self):
     # Iterator needs to work even if message and map are deleted.
@@ -1568,7 +1572,7 @@ class Proto3Test(unittest.TestCase):
     msg.map_int32_int32[3] = 6
     msg.map_int32_int32[4] = 8
 
-    it = msg.map_int32_int32.iteritems()
+    it = msg.map_int32_int32.items()
     del msg
 
     matching_dict = {2: 4, 3: 6, 4: 8}
@@ -1596,7 +1600,7 @@ class Proto3Test(unittest.TestCase):
 
     msg.ClearField('map_int32_int32')
     matching_dict = {2: 4, 3: 6, 4: 8}
-    self.assertMapIterEquals(map.iteritems(), matching_dict)
+    self.assertMapIterEquals(map.items(), matching_dict)
 
   def testMapIterValidAfterFieldCleared(self):
     # Map iterator needs to work even if field is cleared.
@@ -1608,7 +1612,7 @@ class Proto3Test(unittest.TestCase):
     msg.map_int32_int32[3] = 6
     msg.map_int32_int32[4] = 8
 
-    it = msg.map_int32_int32.iteritems()
+    it = msg.map_int32_int32.items()
 
     msg.ClearField('map_int32_int32')
     matching_dict = {2: 4, 3: 6, 4: 8}

python/google/protobuf/internal/proto_builder_test.py

@@ -32,8 +32,15 @@
 
 """Tests for google.protobuf.proto_builder."""
 
-import collections
-import unittest
+try:
+    from collections import OrderedDict
+except ImportError: 
+    from ordereddict import OrderedDict  #PY26
+
+try:
+  import unittest2 as unittest  #PY26
+except ImportError:
+  import unittest
 
 from google.protobuf import descriptor_pb2
 from google.protobuf import descriptor_pool
@@ -44,7 +51,7 @@ from google.protobuf import text_format
 class ProtoBuilderTest(unittest.TestCase):
 
   def setUp(self):
-    self.ordered_fields = collections.OrderedDict([
+    self.ordered_fields = OrderedDict([
         ('foo', descriptor_pb2.FieldDescriptorProto.TYPE_INT64),
         ('bar', descriptor_pb2.FieldDescriptorProto.TYPE_STRING),
         ])

python/google/protobuf/internal/python_message.py

@@ -28,8 +28,6 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-# Keep it Python2.5 compatible for GAE.
-#
 # Copyright 2007 Google Inc. All Rights Reserved.
 #
 # This code is meant to work on Python 2.4 and above only.
@@ -54,21 +52,14 @@ this file*.
 
 __author__ = 'robinson@google.com (Will Robinson)'
 
+from io import BytesIO
 import sys
-if sys.version_info[0] < 3:
-  try:
-    from cStringIO import StringIO as BytesIO
-  except ImportError:
-    from StringIO import StringIO as BytesIO
-  import copy_reg as copyreg
-  _basestring = basestring
-else:
-  from io import BytesIO
-  import copyreg
-  _basestring = str
 import struct
 import weakref
 
+import six
+import six.moves.copyreg as copyreg
+
 # We use "as" to avoid name collisions with variables.
 from google.protobuf.internal import containers
 from google.protobuf.internal import decoder
@@ -355,7 +346,7 @@ def _AttachFieldHelpers(cls, field_descriptor):
 
 def _AddClassAttributesForNestedExtensions(descriptor, dictionary):
   extension_dict = descriptor.extensions_by_name
-  for extension_name, extension_field in extension_dict.iteritems():
+  for extension_name, extension_field in extension_dict.items():
     assert extension_name not in dictionary
     dictionary[extension_name] = extension_field
 
@@ -458,7 +449,7 @@ def _ReraiseTypeErrorWithFieldName(message_name, field_name):
     exc = TypeError('%s for field %s.%s' % (str(exc), message_name, field_name))
 
   # re-raise possibly-amended exception with original traceback:
-  raise type(exc)(exc, sys.exc_info()[2])
+  six.reraise(type(exc), exc, sys.exc_info()[2])
 
 
 def _AddInitMethod(message_descriptor, cls):
@@ -471,7 +462,7 @@ def _AddInitMethod(message_descriptor, cls):
     enum_type with the same name.  If the value is not a string, it's
     returned as-is.  (No conversion or bounds-checking is done.)
     """
-    if isinstance(value, _basestring):
+    if isinstance(value, six.string_types):
       try:
         return enum_type.values_by_name[value].number
       except KeyError:
@@ -493,7 +484,7 @@ def _AddInitMethod(message_descriptor, cls):
     self._is_present_in_parent = False
     self._listener = message_listener_mod.NullMessageListener()
     self._listener_for_children = _Listener(self)
-    for field_name, field_value in kwargs.iteritems():
+    for field_name, field_value in kwargs.items():
       field = _GetFieldByName(message_descriptor, field_name)
       if field is None:
         raise TypeError("%s() got an unexpected keyword argument '%s'" %
@@ -740,7 +731,7 @@ def _AddPropertiesForNonRepeatedCompositeField(field, cls):
 def _AddPropertiesForExtensions(descriptor, cls):
   """Adds properties for all fields in this protocol message type."""
   extension_dict = descriptor.extensions_by_name
-  for extension_name, extension_field in extension_dict.iteritems():
+  for extension_name, extension_field in extension_dict.items():
     constant_name = extension_name.upper() + "_FIELD_NUMBER"
     setattr(cls, constant_name, extension_field.number)
 
@@ -795,7 +786,7 @@ def _AddListFieldsMethod(message_descriptor, cls):
   """Helper for _AddMessageMethods()."""
 
   def ListFields(self):
-    all_fields = [item for item in self._fields.iteritems() if _IsPresent(item)]
+    all_fields = [item for item in self._fields.items() if _IsPresent(item)]
     all_fields.sort(key = lambda item: item[0].number)
     return all_fields
 
@@ -1193,7 +1184,7 @@ def _AddIsInitializedMethod(message_descriptor, cls):
             # ScalarMaps can't have any initialization errors.
             pass
         elif field.label == _FieldDescriptor.LABEL_REPEATED:
-          for i in xrange(len(value)):
+          for i in range(len(value)):
             element = value[i]
             prefix = "%s[%d]." % (name, i)
             sub_errors = element.FindInitializationErrors()
@@ -1223,7 +1214,7 @@ def _AddMergeFromMethod(cls):
 
     fields = self._fields
 
-    for field, value in msg._fields.iteritems():
+    for field, value in msg._fields.items():
       if field.label == LABEL_REPEATED:
         field_value = fields.get(field)
         if field_value is None:

python/google/protobuf/internal/service_reflection_test.py

@@ -34,7 +34,10 @@
 
 __author__ = 'petar@google.com (Petar Petrov)'
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf import unittest_pb2
 from google.protobuf import service_reflection
 from google.protobuf import service
@@ -80,7 +83,7 @@ class FooUnitTest(unittest.TestCase):
     self.assertEqual('Method Bar not implemented.',
                      rpc_controller.failure_message)
     self.assertEqual(None, self.callback_response)
-    
+
     class MyServiceImpl(unittest_pb2.TestService):
       def Foo(self, rpc_controller, request, done):
         self.foo_called = True
@@ -125,8 +128,7 @@ class FooUnitTest(unittest.TestCase):
     # Invoke method.
     stub.Foo(rpc_controller, request, MyCallback)
 
-    self.assertTrue(isinstance(self.callback_response,
-                               unittest_pb2.FooResponse))
+    self.assertIsInstance(self.callback_response, unittest_pb2.FooResponse)
     self.assertEqual(request, channel.request)
     self.assertEqual(rpc_controller, channel.controller)
     self.assertEqual(stub.GetDescriptor().methods[0], channel.method)

python/google/protobuf/internal/symbol_database_test.py

@@ -32,7 +32,10 @@
 
 """Tests for google.protobuf.symbol_database."""
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf import unittest_pb2
 from google.protobuf import symbol_database
 
@@ -64,53 +67,53 @@ class SymbolDatabaseTest(unittest.TestCase):
         messages['protobuf_unittest.TestAllTypes'])
 
   def testGetSymbol(self):
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes, self._Database().GetSymbol(
             'protobuf_unittest.TestAllTypes'))
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes.NestedMessage, self._Database().GetSymbol(
             'protobuf_unittest.TestAllTypes.NestedMessage'))
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes.OptionalGroup, self._Database().GetSymbol(
             'protobuf_unittest.TestAllTypes.OptionalGroup'))
-    self.assertEquals(
+    self.assertEqual(
         unittest_pb2.TestAllTypes.RepeatedGroup, self._Database().GetSymbol(
             'protobuf_unittest.TestAllTypes.RepeatedGroup'))
 
   def testEnums(self):
     # Check registration of types in the pool.
-    self.assertEquals(
+    self.assertEqual(
         'protobuf_unittest.ForeignEnum',
         self._Database().pool.FindEnumTypeByName(
             'protobuf_unittest.ForeignEnum').full_name)
-    self.assertEquals(
+    self.assertEqual(
         'protobuf_unittest.TestAllTypes.NestedEnum',
         self._Database().pool.FindEnumTypeByName(
             'protobuf_unittest.TestAllTypes.NestedEnum').full_name)
 
   def testFindMessageTypeByName(self):
-    self.assertEquals(
+    self.assertEqual(
         'protobuf_unittest.TestAllTypes',
         self._Database().pool.FindMessageTypeByName(
             'protobuf_unittest.TestAllTypes').full_name)
-    self.assertEquals(
+    self.assertEqual(
         'protobuf_unittest.TestAllTypes.NestedMessage',
         self._Database().pool.FindMessageTypeByName(
             'protobuf_unittest.TestAllTypes.NestedMessage').full_name)
 
   def testFindFindContainingSymbol(self):
     # Lookup based on either enum or message.
-    self.assertEquals(
+    self.assertEqual(
         'google/protobuf/unittest.proto',
         self._Database().pool.FindFileContainingSymbol(
             'protobuf_unittest.TestAllTypes.NestedEnum').name)
-    self.assertEquals(
+    self.assertEqual(
         'google/protobuf/unittest.proto',
         self._Database().pool.FindFileContainingSymbol(
             'protobuf_unittest.TestAllTypes').name)
 
   def testFindFileByName(self):
-    self.assertEquals(
+    self.assertEqual(
         'google/protobuf/unittest.proto',
         self._Database().pool.FindFileByName(
             'google/protobuf/unittest.proto').name)

python/google/protobuf/internal/text_encoding_test.py

@@ -32,7 +32,10 @@
 
 """Tests for google.protobuf.text_encoding."""
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf import text_encoding
 
 TEST_VALUES = [
@@ -53,15 +56,15 @@ TEST_VALUES = [
 class TextEncodingTestCase(unittest.TestCase):
   def testCEscape(self):
     for escaped, escaped_utf8, unescaped in TEST_VALUES:
-      self.assertEquals(escaped,
+      self.assertEqual(escaped,
                         text_encoding.CEscape(unescaped, as_utf8=False))
-      self.assertEquals(escaped_utf8,
+      self.assertEqual(escaped_utf8,
                         text_encoding.CEscape(unescaped, as_utf8=True))
 
   def testCUnescape(self):
     for escaped, escaped_utf8, unescaped in TEST_VALUES:
-      self.assertEquals(unescaped, text_encoding.CUnescape(escaped))
-      self.assertEquals(unescaped, text_encoding.CUnescape(escaped_utf8))
+      self.assertEqual(unescaped, text_encoding.CUnescape(escaped))
+      self.assertEqual(unescaped, text_encoding.CUnescape(escaped_utf8))
 
 
 if __name__ == "__main__":

python/google/protobuf/internal/text_format_test.py

@@ -35,10 +35,13 @@
 __author__ = 'kenton@google.com (Kenton Varda)'
 
 import re
+import six
 import string
-import unittest
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf.internal import _parameterized
 
 from google.protobuf import map_unittest_pb2
@@ -62,7 +65,7 @@ class TextFormatBase(unittest.TestCase):
     self.assertMultiLineEqual(text, ''.join(golden_lines))
 
   def CompareToGoldenText(self, text, golden_text):
-    self.assertMultiLineEqual(text, golden_text)
+    self.assertEqual(text, golden_text)
 
   def RemoveRedundantZeros(self, text):
     # Some platforms print 1e+5 as 1e+005.  This is fine, but we need to remove
@@ -101,7 +104,7 @@ class TextFormatTest(TextFormatBase):
         'repeated_string: "\\303\\274\\352\\234\\237"\n')
 
   def testPrintExoticUnicodeSubclass(self, message_module):
-    class UnicodeSub(unicode):
+    class UnicodeSub(six.text_type):
       pass
     message = message_module.TestAllTypes()
     message.repeated_string.append(UnicodeSub(u'\u00fc\ua71f'))
@@ -173,7 +176,7 @@ class TextFormatTest(TextFormatBase):
     parsed_message = message_module.TestAllTypes()
     r = text_format.Parse(wire_text, parsed_message)
     self.assertIs(r, parsed_message)
-    self.assertEquals(message, parsed_message)
+    self.assertEqual(message, parsed_message)
 
     # Test as_utf8 = True.
     wire_text = text_format.MessageToString(
@@ -181,7 +184,7 @@ class TextFormatTest(TextFormatBase):
     parsed_message = message_module.TestAllTypes()
     r = text_format.Parse(wire_text, parsed_message)
     self.assertIs(r, parsed_message)
-    self.assertEquals(message, parsed_message,
+    self.assertEqual(message, parsed_message,
                       '\n%s != %s' % (message, parsed_message))
 
   def testPrintRawUtf8String(self, message_module):
@@ -191,7 +194,7 @@ class TextFormatTest(TextFormatBase):
     self.CompareToGoldenText(text, 'repeated_string: "\303\274\352\234\237"\n')
     parsed_message = message_module.TestAllTypes()
     text_format.Parse(text, parsed_message)
-    self.assertEquals(message, parsed_message,
+    self.assertEqual(message, parsed_message,
                       '\n%s != %s' % (message, parsed_message))
 
   def testPrintFloatFormat(self, message_module):
@@ -218,13 +221,13 @@ class TextFormatTest(TextFormatBase):
     text_message = text_format.MessageToString(message, float_format='.15g')
     self.CompareToGoldenText(
         self.RemoveRedundantZeros(text_message),
-        'payload {{\n  {}\n  {}\n  {}\n  {}\n}}\n'.format(*formatted_fields))
+        'payload {{\n  {0}\n  {1}\n  {2}\n  {3}\n}}\n'.format(*formatted_fields))
     # as_one_line=True is a separate code branch where float_format is passed.
     text_message = text_format.MessageToString(message, as_one_line=True,
                                                float_format='.15g')
     self.CompareToGoldenText(
         self.RemoveRedundantZeros(text_message),
-        'payload {{ {} {} {} {} }}'.format(*formatted_fields))
+        'payload {{ {0} {1} {2} {3} }}'.format(*formatted_fields))
 
   def testMessageToString(self, message_module):
     message = message_module.ForeignMessage()
@@ -287,7 +290,7 @@ class TextFormatTest(TextFormatBase):
     message = message_module.TestAllTypes()
     text = ''
     text_format.Parse(text, message)
-    self.assertEquals(message_module.TestAllTypes(), message)
+    self.assertEqual(message_module.TestAllTypes(), message)
 
   def testParseInvalidUtf8(self, message_module):
     message = message_module.TestAllTypes()
@@ -297,7 +300,7 @@ class TextFormatTest(TextFormatBase):
   def testParseSingleWord(self, message_module):
     message = message_module.TestAllTypes()
     text = 'foo'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         (r'1:1 : Message type "\w+.TestAllTypes" has no field named '
          r'"foo".'),
@@ -306,7 +309,7 @@ class TextFormatTest(TextFormatBase):
   def testParseUnknownField(self, message_module):
     message = message_module.TestAllTypes()
     text = 'unknown_field: 8\n'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         (r'1:1 : Message type "\w+.TestAllTypes" has no field named '
          r'"unknown_field".'),
@@ -315,7 +318,7 @@ class TextFormatTest(TextFormatBase):
   def testParseBadEnumValue(self, message_module):
     message = message_module.TestAllTypes()
     text = 'optional_nested_enum: BARR'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         (r'1:23 : Enum type "\w+.TestAllTypes.NestedEnum" '
          r'has no value named BARR.'),
@@ -323,7 +326,7 @@ class TextFormatTest(TextFormatBase):
 
     message = message_module.TestAllTypes()
     text = 'optional_nested_enum: 100'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         (r'1:23 : Enum type "\w+.TestAllTypes.NestedEnum" '
          r'has no value with number 100.'),
@@ -332,7 +335,7 @@ class TextFormatTest(TextFormatBase):
   def testParseBadIntValue(self, message_module):
     message = message_module.TestAllTypes()
     text = 'optional_int32: bork'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         ('1:17 : Couldn\'t parse integer: bork'),
         text_format.Parse, text, message)
@@ -402,7 +405,7 @@ class OnlyWorksWithProto2RightNowTests(TextFormatBase):
 
     message = unittest_pb2.TestAllTypes()
     test_util.SetAllFields(message)
-    self.assertEquals(message, parsed_message)
+    self.assertEqual(message, parsed_message)
 
   def testPrintAllFields(self):
     message = unittest_pb2.TestAllTypes()
@@ -455,7 +458,7 @@ class OnlyWorksWithProto2RightNowTests(TextFormatBase):
 
     message = unittest_pb2.TestAllTypes()
     test_util.SetAllFields(message)
-    self.assertEquals(message, parsed_message)
+    self.assertEqual(message, parsed_message)
 
   def testPrintMap(self):
     message = map_unittest_pb2.TestMap()
@@ -586,8 +589,8 @@ class Proto2Tests(TextFormatBase):
     text_format.Parse(text, message)
     ext1 = unittest_mset_pb2.TestMessageSetExtension1.message_set_extension
     ext2 = unittest_mset_pb2.TestMessageSetExtension2.message_set_extension
-    self.assertEquals(23, message.message_set.Extensions[ext1].i)
-    self.assertEquals('foo', message.message_set.Extensions[ext2].str)
+    self.assertEqual(23, message.message_set.Extensions[ext1].i)
+    self.assertEqual('foo', message.message_set.Extensions[ext2].str)
 
   def testPrintAllExtensions(self):
     message = unittest_pb2.TestAllExtensions()
@@ -612,7 +615,7 @@ class Proto2Tests(TextFormatBase):
 
     message = unittest_pb2.TestAllExtensions()
     test_util.SetAllExtensions(message)
-    self.assertEquals(message, parsed_message)
+    self.assertEqual(message, parsed_message)
 
   def testParseAllExtensions(self):
     message = unittest_pb2.TestAllExtensions()
@@ -626,12 +629,12 @@ class Proto2Tests(TextFormatBase):
   def testParseBadExtension(self):
     message = unittest_pb2.TestAllExtensions()
     text = '[unknown_extension]: 8\n'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         '1:2 : Extension "unknown_extension" not registered.',
         text_format.Parse, text, message)
     message = unittest_pb2.TestAllTypes()
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         ('1:2 : Message type "protobuf_unittest.TestAllTypes" does not have '
          'extensions.'),
@@ -650,7 +653,7 @@ class Proto2Tests(TextFormatBase):
     message = unittest_pb2.TestAllExtensions()
     text = ('[protobuf_unittest.optional_int32_extension]: 42 '
             '[protobuf_unittest.optional_int32_extension]: 67')
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         ('1:96 : Message type "protobuf_unittest.TestAllExtensions" '
          'should not have multiple '
@@ -661,7 +664,7 @@ class Proto2Tests(TextFormatBase):
     message = unittest_pb2.TestAllTypes()
     text = ('optional_nested_message { bb: 1 } '
             'optional_nested_message { bb: 2 }')
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         ('1:65 : Message type "protobuf_unittest.TestAllTypes.NestedMessage" '
          'should not have multiple "bb" fields.'),
@@ -671,7 +674,7 @@ class Proto2Tests(TextFormatBase):
     message = unittest_pb2.TestAllTypes()
     text = ('optional_int32: 42 '
             'optional_int32: 67')
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError,
         ('1:36 : Message type "protobuf_unittest.TestAllTypes" should not '
          'have multiple "optional_int32" fields.'),
@@ -680,11 +683,11 @@ class Proto2Tests(TextFormatBase):
   def testParseGroupNotClosed(self):
     message = unittest_pb2.TestAllTypes()
     text = 'RepeatedGroup: <'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError, '1:16 : Expected ">".',
         text_format.Parse, text, message)
     text = 'RepeatedGroup: {'
-    self.assertRaisesRegexp(
+    six.assertRaisesRegex(self, 
         text_format.ParseError, '1:16 : Expected "}".',
         text_format.Parse, text, message)
 

python/google/protobuf/internal/type_checkers.py

@@ -28,8 +28,6 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-#PY25 compatible for GAE.
-#
 # Copyright 2008 Google Inc. All Rights Reserved.
 
 """Provides type checking routines.
@@ -49,9 +47,11 @@ TYPE_TO_DESERIALIZE_METHOD: A dictionary with field types and deserialization
 
 __author__ = 'robinson@google.com (Will Robinson)'
 
-import sys  ##PY25
-if sys.version < '2.6': bytes = str  ##PY25
-from google.protobuf.internal import api_implementation
+import six
+
+if six.PY3:
+  long = int
+
 from google.protobuf.internal import decoder
 from google.protobuf.internal import encoder
 from google.protobuf.internal import wire_format
@@ -117,9 +117,9 @@ class IntValueChecker(object):
   """Checker used for integer fields.  Performs type-check and range check."""
 
   def CheckValue(self, proposed_value):
-    if not isinstance(proposed_value, (int, long)):
+    if not isinstance(proposed_value, six.integer_types):
       message = ('%.1024r has type %s, but expected one of: %s' %
-                 (proposed_value, type(proposed_value), (int, long)))
+                 (proposed_value, type(proposed_value), six.integer_types))
       raise TypeError(message)
     if not self._MIN <= proposed_value <= self._MAX:
       raise ValueError('Value out of range: %d' % proposed_value)
@@ -141,9 +141,9 @@ class EnumValueChecker(object):
     self._enum_type = enum_type
 
   def CheckValue(self, proposed_value):
-    if not isinstance(proposed_value, (int, long)):
+    if not isinstance(proposed_value, six.integer_types):
       message = ('%.1024r has type %s, but expected one of: %s' %
-                 (proposed_value, type(proposed_value), (int, long)))
+                 (proposed_value, type(proposed_value), six.integer_types))
       raise TypeError(message)
     if proposed_value not in self._enum_type.values_by_number:
       raise ValueError('Unknown enum value: %d' % proposed_value)
@@ -161,9 +161,9 @@ class UnicodeValueChecker(object):
   """
 
   def CheckValue(self, proposed_value):
-    if not isinstance(proposed_value, (bytes, unicode)):
+    if not isinstance(proposed_value, (bytes, six.text_type)):
       message = ('%.1024r has type %s, but expected one of: %s' %
-                 (proposed_value, type(proposed_value), (bytes, unicode)))
+                 (proposed_value, type(proposed_value), (bytes, six.text_type)))
       raise TypeError(message)
 
     # If the value is of type 'bytes' make sure that it is valid UTF-8 data.

python/google/protobuf/internal/unknown_fields_test.py

@@ -35,7 +35,11 @@
 
 __author__ = 'bohdank@google.com (Bohdan Koval)'
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
+
 from google.protobuf import unittest_mset_pb2
 from google.protobuf import unittest_pb2
 from google.protobuf import unittest_proto3_arena_pb2

python/google/protobuf/internal/wire_format_test.py

@@ -34,7 +34,10 @@
 
 __author__ = 'robinson@google.com (Will Robinson)'
 
-import unittest
+try:
+  import unittest2 as unittest
+except ImportError:
+  import unittest
 from google.protobuf import message
 from google.protobuf.internal import wire_format
 

python/google/protobuf/message_factory.py

@@ -28,8 +28,6 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-#PY25 compatible for GAE.
-#
 # Copyright 2012 Google Inc. All Rights Reserved.
 
 """Provides a factory class for generating dynamic messages.
@@ -43,7 +41,6 @@ my_proto_instance = message_classes['some.proto.package.MessageName']()
 
 __author__ = 'matthewtoia@google.com (Matt Toia)'
 
-import sys  ##PY25
 from google.protobuf import descriptor_database
 from google.protobuf import descriptor_pool
 from google.protobuf import message
@@ -75,8 +72,7 @@ class MessageFactory(object):
     """
     if descriptor.full_name not in self._classes:
       descriptor_name = descriptor.name
-      if sys.version_info[0] < 3:  ##PY25
-##!PY25      if str is bytes:  # PY2
+      if str is bytes:  # PY2
         descriptor_name = descriptor.name.encode('ascii', 'ignore')
       result_class = reflection.GeneratedProtocolMessageType(
           descriptor_name,
@@ -111,7 +107,7 @@ class MessageFactory(object):
     result = {}
     for file_name in files:
       file_desc = self.pool.FindFileByName(file_name)
-      for name, msg in file_desc.message_types_by_name.iteritems():
+      for name, msg in file_desc.message_types_by_name.items():
         if file_desc.package:
           full_name = '.'.join([file_desc.package, name])
         else:
@@ -128,7 +124,7 @@ class MessageFactory(object):
       # ignore the registration if the original was the same, or raise
       # an error if they were different.
 
-      for name, extension in file_desc.extensions_by_name.iteritems():
+      for name, extension in file_desc.extensions_by_name.items():
         if extension.containing_type.full_name not in self._classes:
           self.GetPrototype(extension.containing_type)
         extended_class = self._classes[extension.containing_type.full_name]

python/google/protobuf/proto_builder.py

@@ -30,7 +30,10 @@
 
 """Dynamic Protobuf class creator."""
 
-import collections
+try:
+    from collections import OrderedDict
+except ImportError:
+    from ordereddict import OrderedDict  #PY26
 import hashlib
 import os
 
@@ -80,7 +83,7 @@ def MakeSimpleProtoClass(fields, full_name, pool=None):
   # an OrderedDict we keep the order, but otherwise we sort the field to ensure
   # consistent ordering.
   field_items = fields.items()
-  if not isinstance(fields, collections.OrderedDict):
+  if not isinstance(fields, OrderedDict):
     field_items = sorted(field_items)
 
   # Use a consistent file name that is unlikely to conflict with any imported

python/google/protobuf/text_encoding.py

@@ -27,16 +27,13 @@
 # THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-#PY25 compatible for GAE.
-#
 """Encoding related utilities."""
-
 import re
-import sys  ##PY25
+
+import six
 
 # Lookup table for utf8
-_cescape_utf8_to_str = [chr(i) for i in xrange(0, 256)]
+_cescape_utf8_to_str = [chr(i) for i in range(0, 256)]
 _cescape_utf8_to_str[9] = r'\t'  # optional escape
 _cescape_utf8_to_str[10] = r'\n'  # optional escape
 _cescape_utf8_to_str[13] = r'\r'  # optional escape
@@ -46,9 +43,9 @@ _cescape_utf8_to_str[34] = r'\"'  # necessary escape
 _cescape_utf8_to_str[92] = r'\\'  # necessary escape
 
 # Lookup table for non-utf8, with necessary escapes at (o >= 127 or o < 32)
-_cescape_byte_to_str = ([r'\%03o' % i for i in xrange(0, 32)] +
-                        [chr(i) for i in xrange(32, 127)] +
-                        [r'\%03o' % i for i in xrange(127, 256)])
+_cescape_byte_to_str = ([r'\%03o' % i for i in range(0, 32)] +
+                        [chr(i) for i in range(32, 127)] +
+                        [r'\%03o' % i for i in range(127, 256)])
 _cescape_byte_to_str[9] = r'\t'  # optional escape
 _cescape_byte_to_str[10] = r'\n'  # optional escape
 _cescape_byte_to_str[13] = r'\r'  # optional escape
@@ -75,7 +72,7 @@ def CEscape(text, as_utf8):
   """
   # PY3 hack: make Ord work for str and bytes:
   # //platforms/networking/data uses unicode here, hence basestring.
-  Ord = ord if isinstance(text, basestring) else lambda x: x
+  Ord = ord if isinstance(text, six.string_types) else lambda x: x
   if as_utf8:
     return ''.join(_cescape_utf8_to_str[Ord(c)] for c in text)
   return ''.join(_cescape_byte_to_str[Ord(c)] for c in text)
@@ -100,8 +97,7 @@ def CUnescape(text):
   # allow single-digit hex escapes (like '\xf').
   result = _CUNESCAPE_HEX.sub(ReplaceHex, text)
 
-  if sys.version_info[0] < 3:  ##PY25
-##!PY25  if str is bytes:  # PY2
+  if str is bytes:  # PY2
     return result.decode('string_escape')
   result = ''.join(_cescape_highbit_to_str[ord(c)] for c in result)
   return (result.encode('ascii')  # Make it bytes to allow decode.

python/google/protobuf/text_format.py

@@ -28,17 +28,20 @@
 # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-#PY25 compatible for GAE.
-#
 # Copyright 2007 Google Inc. All Rights Reserved.
 
 """Contains routines for printing protocol messages in text format."""
 
 __author__ = 'kenton@google.com (Kenton Varda)'
 
-import cStringIO
+import io
 import re
 
+import six
+
+if six.PY3:
+  long = int
+
 from google.protobuf.internal import type_checkers
 from google.protobuf import descriptor
 from google.protobuf import text_encoding
@@ -89,7 +92,7 @@ def MessageToString(message, as_utf8=False, as_one_line=False,
   Returns:
     A string of the text formatted protocol buffer message.
   """
-  out = cStringIO.StringIO()
+  out = io.BytesIO()
   PrintMessage(message, out, as_utf8=as_utf8, as_one_line=as_one_line,
                pointy_brackets=pointy_brackets,
                use_index_order=use_index_order,
@@ -136,7 +139,6 @@ def PrintMessage(message, out, indent=0, as_utf8=False, as_one_line=False,
                  use_index_order=use_index_order,
                  float_format=float_format)
 
-
 def PrintField(field, value, out, indent=0, as_utf8=False, as_one_line=False,
                pointy_brackets=False, use_index_order=False, float_format=None):
   """Print a single field name/value pair.  For repeated fields, the value
@@ -157,7 +159,11 @@ def PrintField(field, value, out, indent=0, as_utf8=False, as_one_line=False,
     # For groups, use the capitalized name.
     out.write(field.message_type.name)
   else:
-    out.write(field.name)
+    if isinstance(field.name, six.text_type):
+      name = field.name.encode('utf-8')
+    else:
+      name = field.name
+    out.write(name)
 
   if field.cpp_type != descriptor.FieldDescriptor.CPPTYPE_MESSAGE:
     # The colon is optional in this case, but our cross-language golden files
@@ -211,7 +217,7 @@ def PrintFieldValue(field, value, out, indent=0, as_utf8=False,
       out.write(str(value))
   elif field.cpp_type == descriptor.FieldDescriptor.CPPTYPE_STRING:
     out.write('\"')
-    if isinstance(value, unicode):
+    if isinstance(value, six.text_type):
       out_value = value.encode('utf-8')
     else:
       out_value = value
@@ -537,7 +543,7 @@ class _Tokenizer(object):
   def _PopLine(self):
     while len(self._current_line) <= self._column:
       try:
-        self._current_line = self._lines.next()
+        self._current_line = next(self._lines)
       except StopIteration:
         self._current_line = ''
         self._more_lines = False
@@ -607,7 +613,7 @@ class _Tokenizer(object):
     """
     try:
       result = ParseInteger(self.token, is_signed=True, is_long=False)
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result
@@ -623,7 +629,7 @@ class _Tokenizer(object):
     """
     try:
       result = ParseInteger(self.token, is_signed=False, is_long=False)
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result
@@ -639,7 +645,7 @@ class _Tokenizer(object):
     """
     try:
       result = ParseInteger(self.token, is_signed=True, is_long=True)
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result
@@ -655,7 +661,7 @@ class _Tokenizer(object):
     """
     try:
       result = ParseInteger(self.token, is_signed=False, is_long=True)
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result
@@ -671,7 +677,7 @@ class _Tokenizer(object):
     """
     try:
       result = ParseFloat(self.token)
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result
@@ -687,7 +693,7 @@ class _Tokenizer(object):
     """
     try:
       result = ParseBool(self.token)
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result
@@ -703,8 +709,8 @@ class _Tokenizer(object):
     """
     the_bytes = self.ConsumeByteString()
     try:
-      return unicode(the_bytes, 'utf-8')
-    except UnicodeDecodeError, e:
+      return six.text_type(the_bytes, 'utf-8')
+    except UnicodeDecodeError as e:
       raise self._StringParseError(e)
 
   def ConsumeByteString(self):
@@ -719,8 +725,7 @@ class _Tokenizer(object):
     the_list = [self._ConsumeSingleByteString()]
     while self.token and self.token[0] in ('\'', '"'):
       the_list.append(self._ConsumeSingleByteString())
-    return ''.encode('latin1').join(the_list)  ##PY25
-##!PY25    return b''.join(the_list)
+    return b''.join(the_list)
 
   def _ConsumeSingleByteString(self):
     """Consume one token of a string literal.
@@ -741,7 +746,7 @@ class _Tokenizer(object):
 
     try:
       result = text_encoding.CUnescape(text[1:-1])
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result
@@ -749,7 +754,7 @@ class _Tokenizer(object):
   def ConsumeEnum(self, field):
     try:
       result = ParseEnum(field, self.token)
-    except ValueError, e:
+    except ValueError as e:
       raise self._ParseError(str(e))
     self.NextToken()
     return result

python/setup.py

@@ -8,19 +8,7 @@ import sys
 
 # We must use setuptools, not distutils, because we need to use the
 # namespace_packages option for the "google" package.
-try:
-  from setuptools import setup, Extension, find_packages
-except ImportError:
-  try:
-    from ez_setup import use_setuptools
-    use_setuptools()
-    from setuptools import setup, Extension, find_packages
-  except ImportError:
-    sys.stderr.write(
-        "Could not import setuptools; make sure you have setuptools or "
-        "ez_setup installed.\n"
-    )
-    raise
+from setuptools import setup, Extension, find_packages
 
 from distutils.command.clean import clean as _clean
 
@@ -79,16 +67,14 @@ def generate_proto(source, require = True):
 
     if protoc is None:
       sys.stderr.write(
-          "protoc is not installed nor found in ../src. "
-          "Please compile it or install the binary package.\n"
-      )
+          "protoc is not installed nor found in ../src.  Please compile it "
+          "or install the binary package.\n")
       sys.exit(-1)
 
-    protoc_command = [protoc, "-I../src", "-I.", "--python_out=.", source]
+    protoc_command = [ protoc, "-I../src", "-I.", "--python_out=.", source ]
     if subprocess.call(protoc_command) != 0:
       sys.exit(-1)
 
-
 def GenerateUnittestProtos():
   generate_proto("../src/google/protobuf/map_unittest.proto", False)
   generate_proto("../src/google/protobuf/unittest_arena.proto", False)
@@ -125,13 +111,12 @@ class clean(_clean):
       for filename in filenames:
         filepath = os.path.join(dirpath, filename)
         if filepath.endswith("_pb2.py") or filepath.endswith(".pyc") or \
-           filepath.endswith(".so") or filepath.endswith(".o") or \
-           filepath.endswith('google/protobuf/compiler/__init__.py'):
+          filepath.endswith(".so") or filepath.endswith(".o") or \
+          filepath.endswith('google/protobuf/compiler/__init__.py'):
           os.remove(filepath)
     # _clean is an old-style class, so super() doesn't work.
     _clean.run(self)
 
-
 class build_py(_build_py):
   def run(self):
     # Generate necessary .proto file if it doesn't exist.
@@ -147,13 +132,7 @@ class build_py(_build_py):
         pass
     # _build_py is an old-style class, so super() doesn't work.
     _build_py.run(self)
-  # TODO(mrovner): Subclass to run 2to3 on some files only.
-  # Tracing what https://wiki.python.org/moin/PortingPythonToPy3k's
-  # "Approach 2" section on how to get 2to3 to run on source files during
-  # install under Python 3.  This class seems like a good place to put logic
-  # that calls python3's distutils.util.run_2to3 on the subset of the files we
-  # have in our release that are subject to conversion.
-  # See code reference in previous code review.
+
 
 if __name__ == '__main__':
   ext_module_list = []
@@ -173,6 +152,12 @@ if __name__ == '__main__':
     )
     os.environ['PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION'] = 'cpp'
 
+  # Keep this list of dependencies in sync with tox.ini.
+  install_requires = ['six', 'setuptools']
+  if sys.version_info <= (2,7):
+    install_requires.append('ordereddict')
+    install_requires.append('unittest2')
+
   setup(
       name='protobuf',
       version=GetVersion(),
@@ -183,8 +168,14 @@ if __name__ == '__main__':
       maintainer_email='protobuf@googlegroups.com',
       license='New BSD License',
       classifiers=[
-          'Programming Language :: Python :: 2.7',
-      ],
+        "Programming Language :: Python",
+        "Programming Language :: Python :: 2",
+        "Programming Language :: Python :: 2.6",
+        "Programming Language :: Python :: 2.7",
+        "Programming Language :: Python :: 3",
+        "Programming Language :: Python :: 3.3",
+        "Programming Language :: Python :: 3.4",
+        ],
       namespace_packages=['google'],
       packages=find_packages(
           exclude=[
@@ -196,6 +187,6 @@ if __name__ == '__main__':
           'clean': clean,
           'build_py': build_py,
       },
-      install_requires=['setuptools'],
+      install_requires=install_requires,
       ext_modules=ext_module_list,
   )

python/tox.ini

+[tox]
+envlist =
+    # Py3 tests currently fail because of text handling issues,
+    # So only test py26/py27 for now.
+    #py{26,27,33,34}-{cpp,python}
+    py{26,27}-{cpp,python}
+
+[testenv]
+usedevelop=true
+setenv =
+    cpp: LD_LIBRARY_PATH={toxinidir}/../src/.libs
+    cpp: DYLD_LIBRARY_PATH={toxinidir}/../src/.libs
+    cpp: PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=cpp
+commands =
+    python setup.py -q build_py
+    python: python setup.py -q build
+    cpp: python setup.py -q build --cpp_implementation
+    python: python setup.py -q test -q
+    cpp: python setup.py -q test -q --cpp_implementation
+deps =
+    # Keep this list of dependencies in sync with setup.py.
+    six
+    py26: ordereddict
+    py26: unittest2

src/google/protobuf/compiler/csharp/csharp_helpers.cc

@@ -118,6 +118,12 @@ std::string GetFileNamespace(const FileDescriptor* descriptor) {
 }
 
 std::string GetUmbrellaClassUnqualifiedName(const FileDescriptor* descriptor) {
+  // We manually rename Descriptor to DescriptorProtoFile to avoid collisions with
+  // the static Descriptor property. It would be nice to be able to do this with an
+  // option, but it would be rarely used.
+  if (IsDescriptorProto(descriptor)) {
+    return "DescriptorProtoFile";
+  }
   // umbrella_classname can no longer be set using message option.
   std::string proto_file = descriptor->name();
   int lastslash = proto_file.find_last_of("/");

src/google/protobuf/compiler/csharp/csharp_helpers.h

@@ -115,12 +115,7 @@ inline bool IsMapEntryMessage(const Descriptor* descriptor) {
 // for use in the runtime. This is the only type which is allowed to use proto2 syntax,
 // and it generates internal classes.
 inline bool IsDescriptorProto(const FileDescriptor* descriptor) {
-  // TODO: Do this better! (Currently this depends on a hack in generate_protos.sh to rename
-  // the file...)
-  // We need to be able to detect the "normal" name as well, for times that we're just
-  // depending on descriptor.proto instead of generating it.
-  return descriptor->name() == "google/protobuf/descriptor_proto_file.proto"
-      || descriptor->name() == "google/protobuf/descriptor.proto";
+  return descriptor->name() == "google/protobuf/descriptor.proto";
 }
 
 inline bool IsWrapperType(const FieldDescriptor* descriptor) {

src/google/protobuf/compiler/python/python_generator.cc

@@ -28,7 +28,6 @@
 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-//#PY25 compatible generated code for GAE.
 // Copyright 2007 Google Inc. All Rights Reserved.
 // Author: robinson@google.com (Will Robinson)
 //
@@ -167,7 +166,6 @@ void PrintTopBoilerplate(
   printer->Print(
       "# Generated by the protocol buffer compiler.  DO NOT EDIT!\n"
       "# source: $filename$\n"
-      "\nimport sys\n_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))"  //##PY25
       "\n",
       "filename", file->name());
   if (HasTopLevelEnums(file)) {
@@ -259,12 +257,9 @@ string StringifyDefaultValue(const FieldDescriptor& field) {
     case FieldDescriptor::CPPTYPE_ENUM:
       return SimpleItoa(field.default_value_enum()->number());
     case FieldDescriptor::CPPTYPE_STRING:
-//##!PY25      return "b\"" + CEscape(field.default_value_string()) +
-//##!PY25             (field.type() != FieldDescriptor::TYPE_STRING ? "\"" :
-//##!PY25               "\".decode('utf-8')");
-      return "_b(\"" + CEscape(field.default_value_string()) +  //##PY25
-             (field.type() != FieldDescriptor::TYPE_STRING ? "\")" :  //##PY25
-               "\").decode('utf-8')");  //##PY25
+      return "b\"" + CEscape(field.default_value_string()) +
+             (field.type() != FieldDescriptor::TYPE_STRING ? "\"" :
+               "\".decode('utf-8')");
     case FieldDescriptor::CPPTYPE_MESSAGE:
       return "None";
   }
@@ -390,8 +385,7 @@ void Generator::PrintFileDescriptor() const {
   printer_->Print(m, file_descriptor_template);
   printer_->Indent();
   printer_->Print(
-//##!PY25      "serialized_pb=b'$value$'\n",
-      "serialized_pb=_b('$value$')\n",  //##PY25
+      "serialized_pb=b'$value$'\n",
       "value", strings::CHexEscape(file_descriptor_serialized_));
   if (file_->dependency_count() != 0) {
     printer_->Print(",\ndependencies=[");
@@ -1035,10 +1029,8 @@ string Generator::OptionsValue(
     return "None";
   } else {
     string full_class_name = "descriptor_pb2." + class_name;
-//##!PY25    return "_descriptor._ParseOptions(" + full_class_name + "(), b'"
-//##!PY25        + CEscape(serialized_options)+ "')";
-    return "_descriptor._ParseOptions(" + full_class_name + "(), _b('"  //##PY25
-        + CEscape(serialized_options)+ "'))";  //##PY25
+    return "_descriptor._ParseOptions(" + full_class_name + "(), b'"
+        + CEscape(serialized_options)+ "')";
   }
 }
 

src/google/protobuf/message.cc

@@ -55,6 +55,7 @@
 #include <google/protobuf/stubs/map_util.h>
 #include <google/protobuf/stubs/singleton.h>
 #include <google/protobuf/stubs/stl_util.h>
+#include <google/protobuf/stubs/port.h>
 
 namespace google {
 namespace protobuf {
@@ -485,6 +486,10 @@ struct ShutdownRepeatedFieldRegister {
 
 namespace internal {
 template<>
+#if defined(_MSC_VER) && (_MSC_VER >= 1900)
+// Note: force noinline to workaround MSVC 2015 compiler bug, issue #240
+GOOGLE_ATTRIBUTE_NOINLINE
+#endif
 Message* GenericTypeHandler<Message>::NewFromPrototype(
     const Message* prototype, google::protobuf::Arena* arena) {
   return prototype->New(arena);

src/google/protobuf/stubs/port.h

@@ -153,6 +153,9 @@ static const uint64 kuint64max = GOOGLE_ULONGLONG(0xFFFFFFFFFFFFFFFF);
 // For functions we want to force not inline.
 // Introduced in gcc 3.1.
 #define GOOGLE_ATTRIBUTE_NOINLINE __attribute__ ((noinline))
+#elif defined(_MSC_VER) && (_MSC_VER >= 1400)
+// Seems to have been around since at least Visual Studio 2005
+#define GOOGLE_ATTRIBUTE_NOINLINE __declspec(noinline)
 #else
 // Other compilers will have to figure it out for themselves.
 #define GOOGLE_ATTRIBUTE_NOINLINE

travis.sh

@@ -111,25 +111,45 @@ build_javanano_oracle7() {
   build_javanano
 }
 
+internal_install_python_deps() {
+  sudo pip install tox
+  # Only install Python2.6 on Linux.
+  if [ $(uname -s) == "Linux" ]; then
+    sudo apt-get install -y python-software-properties # for apt-add-repository
+    sudo apt-add-repository -y ppa:fkrull/deadsnakes
+    sudo apt-get update -qq
+    sudo apt-get install -y python2.6 python2.6-dev
+  fi
+}
+
+
 build_python() {
   internal_build_cpp
+  internal_install_python_deps
   cd python
-  python setup.py build
-  python setup.py test
-  python setup.py sdist
-  sudo pip install virtualenv && virtualenv /tmp/protoenv && /tmp/protoenv/bin/pip install dist/*
+  # Only test Python 2.6 on Linux
+  if [ $(uname -s) == "Linux" ]; then
+    envlist=py26-python,py27-python
+  else
+    envlist=py27-python
+  fi
+  tox -e $envlist
   cd ..
 }
 
 build_python_cpp() {
   internal_build_cpp
-  export   LD_LIBRARY_PATH=../src/.libs # for Linux
+  internal_install_python_deps
+  export LD_LIBRARY_PATH=../src/.libs # for Linux
   export DYLD_LIBRARY_PATH=../src/.libs # for OS X
   cd python
-  python setup.py build --cpp_implementation
-  python setup.py test --cpp_implementation
-  python setup.py sdist --cpp_implementation
-  sudo pip install virtualenv && virtualenv /tmp/protoenv && /tmp/protoenv/bin/pip install dist/*
+  # Only test Python 2.6 on Linux
+  if [ $(uname -s) == "Linux" ]; then
+    envlist=py26-cpp,py27-cpp
+  else
+    envlist=py27-cpp
+  fi
+  tox -e $envlist
   cd ..
 }