--- ./doc/tutorial/using_gpu.txt.orig	2016-03-21 13:17:57.000000000 -0600
+++ ./doc/tutorial/using_gpu.txt	2016-03-22 18:49:34.589346869 -0600
@@ -708,7 +708,7 @@ Modify and execute to work for a matrix
 
 
 
-.. _pyCUDA_theano:
+.. _pyCUDA_theano_example:
 
 **Example: Theano + PyCUDA**
 
--- ./MANIFEST.in.orig	2016-03-21 13:17:57.000000000 -0600
+++ ./MANIFEST.in	2016-03-22 18:49:50.492072870 -0600
@@ -6,7 +6,7 @@ global-include *.cpp
 global-include *.h
 global-include *.sh
 global-include *.pkl
-recursive-include docs
+recursive-include docs *
 include bin/theano-cache
 include bin/theano-nose
 include bin/theano-test
--- ./theano/sandbox/cuda/basic_ops.py.orig	2016-03-21 13:17:57.000000000 -0600
+++ ./theano/sandbox/cuda/basic_ops.py	2016-03-22 18:50:25.402276144 -0600
@@ -1279,8 +1279,8 @@ class GpuCAReduce(GpuOp):
         Parameters
         ----------
         N : int
-            The number of 1 in the pattern
-            N=1 -> 01, N=2 -> 011 N=3 ->0111
+            The number of 1s in the pattern.
+            N=1 -> 01, N=2 -> 011, N=3 ->0111.
             Works for N=1,2,3.
 
         """