Fix test for when floatX=float32

Borrow=True is dangerous if one output destroys another

Re-add part of the dtype constraint on out grads

Batched dot22

This is a temporary fix until we fix the behaviour of borrow=True
for the general case.

Re-add (and update) bug fix description.

Err msg unalign

Doc python mem

C version of view op

Fixes after code review

INTERFACE CHANGE: Disable using unaligned tensor in all case. We supported it for python code before.

Added Python 3 support to setup.py

Added Python 3 support to setup.py.  When running under Python 3, setup.py build will now invoke 2to3.

bug scan

In order to avoid expanding memory usage and computations in the part of
the graph that computes gradients, I propose the following conventions,
that re-instate some of the constraint that existed before on the
dtype of gradients:
- When calling some_op.grad(inputs, output_grads), each variable in the
  "output_grads" list, if it is an actual numeric variable (and not,
  for instance, DisconnectedType or NullType), should have the same
  dtype as the corresponding output variable.
- Moreover, if one of the output variables is of a discrete dtype (int
  or uint), then the corresponding output gradient (if not a special
  case like NullType) should be zeros.

This is implemented in theano.grad, so the Op's grad method does not
have to be changed, but now it can rely again on the fact that, if an
output gradient has a dtype, that dtype will be the same as the
corresponding output variable.