fix bytes iteration behaviour: Cython should be consistent whether it knows...

fix bytes iteration behaviour: Cython should be consistent whether it knows the iterable type or not

Cython/Compiler/Optimize.py

@@ -248,6 +248,11 @@ class IterationTransform(Visitor.VisitorTransform):
             unpack_func_type = self.PyUnicode_AS_UNICODE_func_type
             len_func_type = self.PyUnicode_GET_SIZE_func_type
         elif slice_node.type is Builtin.bytes_type:
+            target_type = node.target.type
+            if not target_type.is_int:
+                # bytes iteration returns bytes objects in Py2, but
+                # integers in Py3
+                return node
             unpack_func = "PyBytes_AS_STRING"
             unpack_func_type = self.PyBytes_AS_STRING_func_type
             len_func = "PyBytes_GET_SIZE"

tests/run/for_in_string.pyx

@@ -49,21 +49,43 @@ def for_char_in_bytes(bytes s):
     else:
         return 'X'
 
+#### Py2 and Py3 behave differently here: Py2->bytes, Py3->integer
+##
+## @cython.test_assert_path_exists("//ForFromStatNode")
+## @cython.test_fail_if_path_exists("//ForInStatNode")
+## def for_obj_in_bytes_slice(bytes s):
+##     """
+##     >>> for_obj_in_bytes_slice(bytes_abc)
+##     'X'
+##     >>> for_obj_in_bytes_slice(bytes_ABC)
+##     'B'
+##     >>> for_obj_in_bytes_slice(bytes_abc_null)
+##     'X'
+##     >>> for_obj_in_bytes_slice(bytes_ABC_null)
+##     'B'
+##     """
+##     for c in s[1:-1]:
+##         if c == b'B':
+##             return 'B'
+##     else:
+##         return 'X'
+
 @cython.test_assert_path_exists("//ForFromStatNode")
 @cython.test_fail_if_path_exists("//ForInStatNode")
-def for_obj_in_bytes_slice(bytes s):
+def for_char_in_bytes_slice(bytes s):
     """
-    >>> for_obj_in_bytes_slice(bytes_abc)
+    >>> for_char_in_bytes_slice(bytes_abc)
     'X'
-    >>> for_obj_in_bytes_slice(bytes_ABC)
+    >>> for_char_in_bytes_slice(bytes_ABC)
     'B'
-    >>> for_obj_in_bytes_slice(bytes_abc_null)
+    >>> for_char_in_bytes_slice(bytes_abc_null)
     'X'
-    >>> for_obj_in_bytes_slice(bytes_ABC_null)
+    >>> for_char_in_bytes_slice(bytes_ABC_null)
     'B'
     """
+    cdef char c
     for c in s[1:-1]:
-        if c == b'B':
+        if c == c'B':
             return 'B'
     else:
         return 'X'
@@ -89,20 +111,22 @@ def for_char_in_enumerate_bytes(bytes s):
     else:
         return 'X'
 
-@cython.test_assert_path_exists("//ForFromStatNode")
-@cython.test_fail_if_path_exists("//ForInStatNode")
-def for_pyvar_in_char_ptr(char* c_string):
-    """
-    >>> for_pyvar_in_char_ptr( (bytes_abc+bytes_ABC) * 2 )
-    [True, True, True, False, False, False, True, True, True, False]
-    >>> for_pyvar_in_char_ptr( bytes_abc_null * 2 )
-    [True, False, True, False, True, True, False, True, False, True]
-    """
-    in_test = []
-    cdef object c
-    for c in c_string[:10]:
-        in_test.append( c in b'abc' )
-    return in_test
+#### Py2 and Py3 behave differently here: Py2->bytes, Py3->integer
+##
+## @cython.test_assert_path_exists("//ForFromStatNode")
+## @cython.test_fail_if_path_exists("//ForInStatNode")
+## def for_pyvar_in_char_ptr(char* c_string):
+##     """
+##     >>> for_pyvar_in_char_ptr( (bytes_abc+bytes_ABC) * 2 )
+##     [True, True, True, False, False, False, True, True, True, False]
+##     >>> for_pyvar_in_char_ptr( bytes_abc_null * 2 )
+##     [True, False, True, False, True, True, False, True, False, True]
+##     """
+##     in_test = []
+##     cdef object c
+##     for c in c_string[:10]:
+##         in_test.append( c in b'abc' )
+##     return in_test
 
 @cython.test_assert_path_exists("//ForFromStatNode")
 @cython.test_fail_if_path_exists("//ForInStatNode")