Pull in the development version of zodbpickle which has recent changes which allow tests to pass at this time.

The close() method is not in the interface and is not provided
by ZEO's implementation.

This uses bytes_as_strings=True option introduced in zodbpickle 0.2 for
this purpose.

This way pickles produced on Python 3 are nearly the same as on Python 2.
There are some slight differences (Python 3 seems to perform more
memoizations which grows the size of some pickles by a couple of bytes),
but they're immaterial.

Now we can use zodbpickle's noload() on Python 3 to scan pickles for
persistent references.  We couldn't do that before, because Python 3
normally pickles byte strings as calls to codecs.encode(u'latin1-data',
'latin-1'), and noload() doesn't interpret the REDUCE opcode involved in
that representation.

Note that when you're pickling byte strings using bytes_as_strings=True,
you have to load them using encoding='bytes' (which breaks instances, so
cannot be used) or using errors='bytes' (which mean some bytestrings may
get unpickled as unicode instead).  I've tried hard to discover every
place that unpickles OIDs and added conversion to bytes in those places.

Applications dealing with binary data be prepared to handle bytestrings
that unexpectedly become unicode on unpickling.  That's the price of
Python 2 compatibility.

Instead, just remove the non-ASCII characters, regaining Python 3.2 compat.

Fixes some more of those ResourceWarnings

Many more remain.

Since now we need to pass an extra argument to loads() and Unpickler(),
but only on Python 3, the old way of ``from ZODB._compat import pickle``
no longer pays out, and I had to import the names directly.

The only remaining test failures are due to changed pickle sizes.  I've
discovered two reasons for that so far:

  1) Python 3 pickles 'native strings' as BINUNICODE, Python 2 pickles
     'native strings' as SHORT_BINSTRING, which is 3 bytes shorter.

  2) Python 3 and also the 'pickle' module Python 2 pickles tuples like
     (UserDefinedClass, ) with an extra BINPUT opcode at the end,
     compared to 'cPickle' in Python 2 when given the same tuple:

        Python 2.7.3 (default, Sep 26 2012, 21:51:14)
        [GCC 4.7.2] on linux2
        Type "help", "copyright", "credits" or "license" for more information.
        >>> import pickle, cPickle
        >>> class MyClass(object): pass
        ...
        >>> len(pickle.dumps((MyClass, None), 1))
        26
        >>> len(cPickle.dumps((MyClass, None), 1))
        24

The plan is to switch to zodbpickle instead of the builtin Python 3
pickle, then use encoding=bytes from http://bugs.python.org/issue6784
and fix the class-in-tuple encoding to match Python 2's cPickle.

This test wants to rely on OIDs that are assigned to various objects.
OIDs are assigned by Pickler.persistent_id(), which gets called during
graph traversal in a nonspecific order (courtesy randomized hashes in
Python 3), so if we want to be sure to have the right OIDs assigned to
the right objects, we need to call connection.add() manually or commit
after adding every object.