Fix bug and refactor in cython backend

theano/scan_module/scan_perform.pyx

@@ -62,7 +62,7 @@ import copy
 
 
 def get_version():
-    return 0.288
+    return 0.289
 
 @cython.boundscheck(False)
 def perform(
@@ -445,7 +445,7 @@ def perform(
                     # modified inplace and nothing needs to be done.
                     if not same_data:
                         outs[j][0][<unsigned int>(k + pos[j])] = \
-                            input_storage[<unsigned int>inp_idx].storage[0]
+                            input_storage[<unsigned int>(n_seqs + inp_idx)].storage[0]
 
                 else:
                     # This output tap has not been preallocated, recover
@@ -465,6 +465,10 @@ def perform(
 
         for j in range(begin, end):
 
+            # Copy the output value to `outs`, if necessary
+            if store_steps[j] == 1 or vector_outs[j] == 1:
+                outs[j][0][pos[j]] = output_storage[<unsigned int>(offset_out+j)].storage[0]
+            else:
                 # Check whether the initialization of the output storage map
                 # for this output has been reused.
                 old_var = old_output_storage[offset_out + j]
@@ -480,15 +484,33 @@ def perform(
                 else:
                     output_reused = False
 
-            # Copy the output value to `outs`, if necessary
-            if store_steps[j] == 1 or vector_outs[j] == 1 or not output_reused:
-                outs[j][0][pos[j]] = output_storage[<unsigned int>(offset_out+j)].storage[0]
+                if not output_reused:
+                    outs[j][0][pos[j]] = \
+                        output_storage[<unsigned int>(offset_out+j)].storage[0]
+
 
         # 5.5 Copy over the values for nit_sot outputs
         begin  = end
         end   += n_nit_sot
         for j in range(begin,end):
 
+            if i == 0:
+                jout = j+offset_out
+                shape = (store_steps[j],) + output_storage[jout].storage[0].shape
+                if len(output_storage[jout].storage[0].shape) == 0:
+                    vector_outs[j] = 1
+                dtype = output_storage[jout].storage[0].dtype
+                if (outs[j][0] is None or
+                        outs[j][0].shape[0] < store_steps[j] or
+                        outs[j][0].shape[1:] != shape[1:] or
+                        outs[j][0].dtype != dtype ):
+                    outs[j][0] = node.outputs[j].type.value_zeros(shape)
+                elif outs[j][0].shape[0] != store_steps[j]:
+                    outs[j][0] = outs[j][0][:store_steps[j]]
+                outs[j][0][pos[j]] = output_storage[jout].storage[0]
+            elif store_steps[j] == 1 or vector_outs[j] == 1:
+                outs[j][0][pos[j]] = output_storage[j+offset_out].storage[0]
+            else:
                 # Check whether the initialization of the output storage map
                 # for this output has been reused.
                 old_var = old_output_storage[offset_out + j]
@@ -504,22 +526,7 @@ def perform(
                 else:
                     output_reused = False
 
-            if i == 0:
-                jout = j+offset_out
-                shape = (store_steps[j],) + output_storage[jout].storage[0].shape
-                if len(output_storage[jout].storage[0].shape) == 0:
-                    vector_outs[j] = 1
-                dtype = output_storage[jout].storage[0].dtype
-                if (outs[j][0] is None or
-                        outs[j][0].shape[0] < store_steps[j] or
-                        outs[j][0].shape[1:] != shape[1:] or
-                        outs[j][0].dtype != dtype ):
-                    outs[j][0] = node.outputs[j].type.value_zeros(shape)
-                elif outs[j][0].shape[0] != store_steps[j]:
-                    outs[j][0] = outs[j][0][:store_steps[j]]
-                outs[j][0][pos[j]] = output_storage[jout].storage[0]
-            elif (store_steps[j] == 1 or vector_outs[j] == 1 or
-                  not output_reused):
+                if not output_reused:
                     outs[j][0][pos[j]] = output_storage[j+offset_out].storage[0]
 
         # 5.6 Copy over the values for outputs corresponding to shared

theano/scan_module/scan_perform_ext.py

@@ -17,7 +17,7 @@ from theano.gof import cmodule
 _logger = logging.getLogger('theano.scan_module.scan_perform')
 
 
-version = 0.288  # must match constant returned in function get_version()
+version = 0.289  # must match constant returned in function get_version()
 
 need_reload = False