Use a higher pickle protocol (3) for serializing objects on Python 2

Previously protocol 1 was used. This is more efficient for new-style
classes (all persistent objects are new-style), according to the docs,
at the cost of being very slightly less space efficient for old-style
classes.

In tests of a persistent object with two trivial numeric attributes,
the higher protocol was 12 bytes smaller, and serialized and
deserialized 1us faster. Introducing a reference to another new-style
class for a more realistic test made the higher protocol twice as fast
to serialize (20.5 vs 10.3us), almost half the size (215 vs 142
bytes), and it deserialized 30% faster (6.5 vs 4.6us).

On Python 2, this will now allow open ``file`` objects to be
pickled (loading the object will result in a closed file); previously
this would result in a ``TypeError`` (as does under Python 3). We had
tests that you couldn't do that with a BlobFile so I had to update it
to still make that true.

I wouldn't recommend serializing arbitrary open files under Python
2 (for one thing, they can't trivially be deserialized in Python 3),
but I didn't take any steps to prevent it either. Since this hasn't
been possible, there shouldn't be code in the wild that is trying to
do it---and it wouldn't be forward compatible with Python 3 either.

CHANGES.rst

@@ -12,6 +12,17 @@
 - Ensure that the ``HistoricalStorageAdapter`` forwards the ``release`` method to
   its base instance. See `issue 78 <https://github.com/zopefoundation/ZODB/issues/788>`_.
 
+- Use a higher pickle protocol (3) for serializing objects on Python
+  2; previously protocol 1 was used. This is *much* more efficient for
+  new-style classes (all persistent objects are new-style), at the
+  cost of being very slightly less efficient for old-style classes.
+
+  .. note:: On Python 2, this will now allow open ``file`` objects
+			(but **not** open blobs or sockets) to be pickled (loading
+			the object will result in a closed file); previously this
+			would result in a ``TypeError``. Doing so is not
+			recommended as they cannot be loaded in Python 3.
+
 5.2.4 (2017-05-17)
 ==================
 

src/ZODB/_compat.py

@@ -34,7 +34,7 @@
     HIGHEST_PROTOCOL = cPickle.HIGHEST_PROTOCOL
     IMPORT_MAPPING = {}
     NAME_MAPPING = {}
-    _protocol = 1
+    _protocol = 3
     FILESTORAGE_MAGIC = b"FS21"
 else:
     # Python 3.x: can't use stdlib's pickle because

src/ZODB/blob.py

@@ -338,6 +338,16 @@ def close(self):
         self.blob.closed(self)
         super(BlobFile, self).close()
 
+    def __reduce__(self):
+        # Python 3 cannot pickle an open file with any pickle protocol
+        # because of the underlying _io.BufferedReader/Writer object.
+        # Python 2 cannot pickle a file with a protocol < 2, but
+        # protocol 2 *can* pickle an open file; the result of unpickling
+        # is a closed file object.
+        # It's pointless to do that with a blob, so we make sure to
+        # prohibit it on all versions.
+        raise TypeError("Pickling a BlobFile is not allowed")
+
 _pid = str(os.getpid())
 
 def log(msg, level=logging.INFO, subsys=_pid, exc_info=False):