backport to python 2.4

theano/sandbox/neighbourhoods.py

@@ -65,7 +65,10 @@ class NeighbourhoodsFromImages(Op):
         """
         self.n_dims_before = n_dims_before
         self.dims_neighbourhoods = dims_neighbourhoods
-        self.strides = strides if not strides is None else dims_neighbourhoods
+        if not strides is None:                                                               
+            self.strides = strides                                                            
+        else:                                                                                 
+            self.strides = dims_neighbourhoods                                                
         self.ignore_border = ignore_border
 
         self.inverse = inverse

theano/sandbox/rng_mrg.py

@@ -144,8 +144,10 @@ class mrg_uniform_base(Op):
     def __hash__(self):
         return hash(type(self)) ^ hash(self.output_type) ^ hash(self.inplace)
     def __str__(self):
-        return self.__class__.__name__+"{%s,%s}"%(self.output_type,
-                                                  "inplace" if self.inplace else "no_inplace")
+        if self.inplace:
+            s = "inplace"
+        else: s = "no_inplace"
+        return self.__class__.__name__+"{%s,%s}"%(self.output_type,s)
 
     def make_node(self, rstate, size):
         # error checking slightly redundant here, since

theano/tensor/nnet/tests/test_nnet.py

@@ -7,6 +7,7 @@ import theano
 from theano import tensor as T
 from theano import tensor
 from theano import gof
+from theano.gof.python25 import all
 from theano.tests import unittest_tools as utt
 from theano import printing, pprint
 from theano.tensor.nnet import (categorical_crossentropy,
@@ -928,7 +929,7 @@ def test_asymptotic_32():
         assert gxval[0,1] == 0.25
 
 
-class Test_softmax_opt():
+class Test_softmax_opt:
     # Test that expressions of softmax in terms of exponentiated things divided by row sums
     # are replaced by softmax expressions.
 

theano/tests/test_scan.py

@@ -77,7 +77,10 @@ class multiple_outputs_numeric_grad:
         for i in xrange(len(pt)):
             if ndarray_mask[i]:
                 # It is a ndarray that we can tweak
-                _eps = eps if eps else dtype_eps
+		if eps:
+                    _eps = eps
+                else:
+                    _eps = dtype_eps
                 if pt[i].ndim :
                     _g = []
                     # it has several dimensions: