Split Josh's documentation between aliasing.txt and CudaNdarraySharedVariable.set_value docstring note.

test that we work inplace when accessing the internal of theano to replace the value of a shared variable.

For tensor and sparse shared varible, test that we don't work inplace.(not implemented for now)

I am not sure why this is necessary now.  One guess is that the canonicalizer or
something else has been modified to put in identity(), and now it is more
obvious that we need an optimization to take it out again.  I'm hoping that this
improves the slowdown recently observed in convolutional net benchmark.

The modification ensures that that even if numpy is clever enough to not make a
copy on unit-length reductions, the perform() returns a copy.

Corrected DEBUG_MODE and FAST_COMPILE tests for grad of Prod. Added comments on mechanism for this grad.

fix may_share_memory for sparse type as it was done for ndarray type. Add may_share_memory test for CudaNdarray type

Allow the default value of parameter dx,dy of ConvOp.__init__. They are set explicitly to 1 by the conv() fct.