Support for pickling - Githubissues

I noticed that pickling a Proxy class does not always work, depending on what the underlying object is. This is in Python 2.7.

Typical classes, strings, integers, floats work OK.
datetime returns an error like can't pickle datetime objects
Decimal returns a class that defines __slots__ without defining __getstate__ cannot be pickled
dictionaries/lists don't raise an error when pickling, but when unpickled they are empty

from datetime import datetime
from lazy_object_proxy import Proxy
import pickle

p = Proxy(lambda: datetime(2015, 1, 2))
pickled = pickle.dumps(p, -1)
print pickle.loads(pickled)

...venv/lib/python2.7/copy_reg.pyc in _reduce_ex(self, proto)
     68     else:
     69         if base is self.__class__:
---> 70             raise TypeError, "can't pickle %s objects" % base.__name__
     71         state = base(self)
     72     args = (self.__class__, base, state)

TypeError: can't pickle datetime objects

For me. the desired behavior would be to immediately evaluate the wrapped object and pickle that object. After reading a bit on the pickle documentation, I was able to get something to work for all the types I've encountered in my application - basically had to add a __reduce_ex()__ and __reduce()__ method on my Proxy subclass, with a couple of special cases for certain types. (The more important one was reduce_ex as that is called first). If you're interested I could share this code. That said I'm not an expert on Python's pickle functionality and why it seems to be so different for different data types.

Just wanted to bring this up, I'm not sure if this should be something built into the library or left as something for users to implement in their subclasses (maybe with some documentation on how to do this).

ionelmc / python-lazy-object-proxy

Support for pickling #7