fixed issue with the user retrieving results from the outputs he provided

compile.py

@@ -10,9 +10,15 @@ from copy import copy
 class profile_linker:
     def __init__(self, env):
         self.order = env.toposort()
+#         print "digraph unix { size = '6,6'; node [color = lightblue2; style = filled];"
+#         for op in self.order:
+#             for input in op.inputs:
+#                 if input.owner:
+#                     print input.owner.__class__.__name__ + str(abs(id(input.owner))), " -> ", op.__class__.__name__ + str(abs(id(op))), ";"
         self.thunks = [op._perform for op in self.order]
         self.n_calls = 0
         self.times = [0.0 for op in self.order]
+    
     def __call__(self):
         for i, thunk in enumerate(self.thunks):
             start_time = time.time()
@@ -51,13 +57,14 @@ class prog(gof.Prog):
         TODO: think about whether orphan computation should be in this function,
         or in self.__call__()
         """
-        outputs = gof.mark_outputs_as_destroyed(outputs)
+        new_outputs = gof.mark_outputs_as_destroyed(outputs)
         gof.Prog.__init__(self,
                           inputs,
-                          outputs,
+                          new_outputs,
                           optimizer,
                           linker,
                           [])
+        self.outputs = outputs
         self.compute_orphans()
 
     def __call__(self, check_uncomputed = True):

core.py

@@ -193,6 +193,8 @@ class NumpyR(gof.PythonR):
             self.data = UNCOMPUTED
         elif isinstance(value, numpy.ndarray):
             self.data = value
+        elif isinstance(value, PythonR):
+            self.set_value(value.data)
         else:
             self.data = numpy.array(value)
         self.up_to_date = True

opt.py

@@ -34,5 +34,6 @@ optimizer_end = gof.SeqOptimizer([gof.DummyRemover])
 def optimizer(lst):
     seq_opt = gof.SeqOptimizer(optimizer_begin + lst + optimizer_end)
     rval = gof.PythonOpt(gof.MergeOptMerge(seq_opt))
+#    rval = gof.PythonOpt((seq_opt))
     return rval