Remove useless graph of clone. Scan inputs and outputs aren't modified inplace, so multiple scan can share them.

Make DebugMode keep memory layout when possible.

This will still convert a noncontiguous array to contiguous, but will
not convert F to C.  We can make it keep the noncontiguous layout but
it's more involved and can take a *LOT* more of memory.

fix gh-5694 I didn't test it, but it should now skip empty directory …

Fix offset in GpuIncSubtensor

docs: Fix openmp config section

Removed useless printf.

Clarify that the stable version of 0.8.2 and people don't want to use…

Fix error in Advanced indexing with > 2D

- openmp has been disabled by default since https://github.com/Theano/Theano/commit/f1f06bb960f63e098a7ce32de50c8547d2fb7713.
- grammar improvements

Prepare release 0.9.0rc3.

Last before rc3

The predefined constants (enum) are the same with v6

Fixes #5655

Change the default of conv.assert_shape flags to speed up compilation time

Faster compilation and import with cuda backend

Remove old files not used anymore.

A couple of test fixes

Fix problems in daily buildbot (debug mode)

Warn if taking softmax over broadcastable dimension

We already encountered then in ProfileStats.__init__

The gpu back-end may have been imported, but we are not even timing a GPU function