Merge pull request #613 from delallea/minor

NEWS.txt

@@ -10,12 +10,12 @@ Bug fixes
  * In Sparse sandbox, fix the grad of theano.sparse.sandbox.sp.row_scale.
    It did not return the right number of elements. (Frederic B.)
  * set_subtensor(x[int vector], new_value) when moved to the GPU
-   where transformed into inc_subtensor on the GPU. Now we have a slow
-   GPU implementation.
-   Note: set_subtensor(x[slice[,...]], new_value) was working correctly
-   in all case as well as inc_subtensor(*, *).
-   Note2: If your code have this behavior, we print a warning by default.
-   (Frederic B.)
+   was transformed into inc_subtensor on the GPU. Now we have a correct
+   (but slow) GPU implementation.
+   Note 1: set_subtensor(x[slice[,...]], new_value) was working correctly
+   in all cases as well as inc_subtensor(*, *).
+   Note 2: If your code was affected by the incorrect behavior, we now print
+   a warning by default (Frederic B.)
  * Fixed an issue whereby config values were used as default arguments,
    with those defaults then stuck at old values if the config variables were
    changed during program execution. (David W-F)
@@ -65,25 +65,24 @@ New Features
    (Frederic B., Simon McGregor)
  * MRG random now raises an error with a clear message when the passed shape
    contains dimensions with bad value like 0. (Frédéric B. reported by Ian G.)
- * "CudaNdarray[*] = ndarray" work in more case (Frederic B.)
- * "CudaNdarray[*] += ndarray" work in more case (Frederic B.)
+ * "CudaNdarray[*] = ndarray" works in more cases (Frederic B.)
+ * "CudaNdarray[*] += ndarray" works in more cases (Frederic B.)
  * We add dimensions to CudaNdarray to automatically broadcast more frequently.
    (Frederic B.)
- * theano.tensor.argsort that wrap numpy.argsort (Hani Almousli).
+ * theano.tensor.argsort that wraps numpy.argsort (Hani Almousli).
  * New theano flag cmodule.warn_no_version. Default False. If True,
-  will print a warning when compiling one or more Op with c code that
-  can't be cached as there is no c_code_cache_version() function to at
-  least of of those Ops.  (Frederic B.)
- * CPU alloc now always generate c code (Pascal L.)
+   will print a warning when compiling one or more Op with C code that
+   can't be cached because there is no c_code_cache_version() function
+   associated to at least one of those Ops.  (Frederic B.)
+ * CPU alloc now always generate C code (Pascal L.)
  * New Theano flag cmodule.warn_no_version=False. When True, warn when an op
-   with c code is not versionned. This force to recompile it everytimes.
+   with C code is not versioned (which forces to recompile it everytimes).
    (Frédéric B.)
- * Made a few Ops with c code versionned to reduce compilation time.
+ * Made a few Ops with C code versioned to reduce compilation time.
    (Frédéric B, Pascal L.)
-
- * c code reuse preallocated output(only done by Scan) (Pascal L.)
- * gc of intermediate result during theano function call for op with c code
-   (Pascal L.)
+ * C code reuses preallocated outputs (only done by Scan) (Pascal L.)
+ * Garbage collection of intermediate results during Theano function calls
+   for Ops with C code (Pascal L.)
 
 Sparse
  * Implement theano.sparse.mul(sparse1, sparse2) when both inputs don't
@@ -278,7 +277,7 @@ New features:
  * theano.sandbox.cuda.cuda_ndarray.cuda_ndarray.mem_info() return free and total gpu memory (Frederic)
  * Theano flags compiledir_format. Keep the same default as before: compiledir_%(platform)s-%(processor)s-%(python_version)s. (Josh Bleecher Snyder)
      * We also support the "theano_version" substitution.
- * IntDiv c code (faster and allow this elemwise to be fused with other elemwise) (Pascal)
+ * IntDiv C code (faster and allow this elemwise to be fused with other elemwise) (Pascal)
  * Internal filter_variable mechanism in Type. (Pascal, Ian)
     * Ifelse works on sparse.
     * It makes use of gpu shared variable more transparent with theano.function updates and givens parameter.
@@ -347,7 +346,7 @@ Crashes fixed:
  * Fix runtime crash in gemm, dot22. FB
  * Fix on 32bits computer: make sure all shape are int64.(Olivier)
  * Fix to deque on python 2.4 (Olivier)
- * Fix crash when not using c code (or using DebugMode) (not used by
+ * Fix crash when not using C code (or using DebugMode) (not used by
    default) with numpy 1.6*. Numpy has a bug in the reduction code that
    made it crash. (Pascal)
  * Crashes of blas functions (Gemv on CPU; Ger, Gemv and Gemm on GPU)

doc/library/config.txt

@@ -563,13 +563,13 @@ import theano and print the config variable, as in:
 
     Bool value, default: False
 
-    If True, will print a warning when compiling one or more Op with c
-    code that can't be cached as there is no c_code_cache_version()
-    function to at least of of those Ops.
+    If True, will print a warning when compiling one or more Op with C
+    code that can't be cached because there is no c_code_cache_version()
+    function associated to at least one of those Ops.
 
 .. attribute:: config.cmodule.mac_framework_link
 
     Bool value, default: False
 
-    If set to True, breaks certain mac installations with the infamous
+    If set to True, breaks certain MacOS installations with the infamous
     Bus Error.

theano/configdefaults.py

@@ -317,7 +317,7 @@ AddConfigVar('warn.subtensor_merge_bug',
 
 AddConfigVar('warn.gpu_set_subtensor1',
         "Warn if previous versions of Theano (before 0.6) could have given "
-        "incorrect results when moving to the gpu"
+        "incorrect results when moving to the gpu "
         "set_subtensor(x[int vector], new_value)",
         BoolParam(warn_default('0.6')),
         in_c_key=False)

theano/gof/cmodule.py

@@ -27,14 +27,17 @@ from theano.gof.cc import hash_from_code
 import compilelock
 
 from theano.configparser import AddConfigVar, BoolParam
+
 AddConfigVar('cmodule.mac_framework_link',
-        ("If set to True, breaks certain mac installations with the infamous "
-         "Bus Error"),
+        "If set to True, breaks certain MacOS installations with the infamous "
+        "Bus Error",
         BoolParam(False))
+
 AddConfigVar('cmodule.warn_no_version',
-             "If True, will print a warning when compiling one or more Op with"
-             "c code that can't be cached as there is no c_code_cache_version"
-             "() function to at least of of those Ops.",
+             "If True, will print a warning when compiling one or more Op "
+             "with C code that can't be cached because there is no "
+             "c_code_cache_version() function associated to at least one of "
+             "those Ops.",
              BoolParam(False))
 
 
@@ -969,7 +972,7 @@ class ModuleCache(object):
                                    if isinstance(k, theano.Op)]
                             _logger.warning("not all the"
                                 " following op(s) implement"
-                                " c_code_cache_version(). This make them"
+                                " c_code_cache_version(). This makes them"
                                 " recompiled for each process." + str(ops))
 
                         # Map the new module to its KeyData object. Note that
@@ -1437,7 +1440,7 @@ class GCC_compiler(object):
         #TODO: Do not do the dlimport in this function
 
         if include_dirs is None:
-            preargs = []
+            include_dirs = []
         if lib_dirs is None:
             lib_dirs = []
         if libs is None:

theano/sandbox/cuda/basic_ops.py

@@ -1939,10 +1939,10 @@ class GpuAdvancedIncSubtensor1(tensor.AdvancedIncSubtensor1, GpuOp):
 
         return Apply(self, [x_, y_, ilist_], [x_.type()])
 
-    # CudaNdarray_Subscript() don't support Advanced slicing.
-    # But we can't use the parent version that loop on each indices
+    # CudaNdarray_Subscript() doesn't support Advanced slicing.
+    # But we can't use the parent version that loops on each index
     # as we also need to loop when set_instead_of_inc is True and the
-    # parent don't loop in that case.
+    # parent doesn't loop in that case.
     def perform(self, node, inp, out_):
         # TODO opt to make this inplace
         x, y, idx = inp
@@ -1950,7 +1950,7 @@ class GpuAdvancedIncSubtensor1(tensor.AdvancedIncSubtensor1, GpuOp):
         if not self.inplace:
             x = x.copy()
         if self.set_instead_of_inc:
-            # CudaNdarray __setitem__ don't do broadcast nor support
+            # CudaNdarray __setitem__ doesn't do broadcast nor support
             # list of index.
             assert y.ndim <= x.ndim   # Should be guaranteed by `make_node`
             if y.ndim == x.ndim:

theano/sandbox/cuda/cuda_ndarray.cu

@@ -1066,7 +1066,7 @@ CudaNdarray_inplace_elemwise(PyObject* py_self, PyObject * py_other, operator_t
             PyErr_SetString(
                 PyExc_ValueError,
                 "CudaNdarray_inplace_elemwise cannot work inplace on"
-                " un-initialized array when the new value have more then"
+                " un-initialized array when the new value have more than"
                 " 0 or 1 broadcastable dimensions");
             Py_XDECREF(new_other);
             return 0;
@@ -2799,9 +2799,9 @@ int CudaNdarray_CopyFromCudaNdarray(CudaNdarray * self,
     if (self->nd < other->nd)
     {
         PyErr_Format(PyExc_NotImplementedError,
-            "CudaNdarray_CopyFromCudaNdarray: The destination need more or the"
-            " same number of dimensions then the source. Got %d and %d.",
-                     self->nd, other->nd);
+            "CudaNdarray_CopyFromCudaNdarray: The number of dimensions of the "
+            "destination needs to be >= the number of dimensions of the "
+            "source. Got %d and %d.", self->nd, other->nd);
         return -1;
     }
     else if (self->nd != other->nd)