More merging

Cython/Compiler/CmdLine.py

@@ -50,7 +50,7 @@ Options:
 #                   transforms for the same phase will be used in the order they are given.
 
 def bad_usage():
-    print >>sys.stderr, usage
+    sys.stderr.write(usage)
     sys.exit(1)
 
 def parse_command_line(args):
@@ -140,14 +140,14 @@ def parse_command_line(args):
             elif arg.endswith(".o"):
                 options.objects.append(arg)
             else:
-                print >>sys.stderr, \
-                    "cython: %s: Unknown filename suffix" % arg
+                sys.stderr.write(
+                    "cython: %s: Unknown filename suffix\n" % arg)
     if options.objects and len(sources) > 1:
-        print >>sys.stderr, \
-            "cython: Only one source file allowed together with .o files"
+        sys.stderr.write(
+            "cython: Only one source file allowed together with .o files\n")
     if options.use_listing_file and len(sources) > 1:
-        print >>sys.stderr, \
-            "cython: Only one source file allowed when using -o"
+        sys.stderr.write(
+            "cython: Only one source file allowed when using -o\n")
         sys.exit(1)
     if len(sources) == 0 and not options.show_version:
         bad_usage()

Cython/Compiler/ExprNodes.py

@@ -18,28 +18,6 @@ from Cython.Debugging import print_call_chain
 from DebugFlags import debug_disposal_code, debug_temp_alloc, \
     debug_coercion
 
-class EncodedString(unicode):
-    # unicode string subclass to keep track of the original encoding.
-    # 'encoding' is None for unicode strings and the source encoding
-    # otherwise
-    encoding = None
-
-    def byteencode(self):
-        assert self.encoding is not None
-        return self.encode(self.encoding)
-
-    def utf8encode(self):
-        assert self.encoding is None
-        return self.encode("UTF-8")
-
-    def is_unicode(self):
-        return self.encoding is None
-    is_unicode = property(is_unicode)
-
-#    def __eq__(self, other):
-#        return unicode.__eq__(self, other) and \
-#            getattr(other, 'encoding', '') == self.encoding
-
 
 class ExprNode(Node):
     #  subexprs     [string]     Class var holding names of subexpr node attrs
@@ -2251,7 +2229,7 @@ class SequenceNode(ExprNode):
                 self.iterator.py_result()))
         if debug_disposal_code:
             print("UnpackNode.generate_assignment_code:")
-            print("...generating disposal code for %s" % iterator)
+            print("...generating disposal code for %s" % self.iterator)
         self.iterator.generate_disposal_code(code)
 
         code.putln("}")
@@ -2279,7 +2257,7 @@ class TupleNode(SequenceNode):
             
     def calculate_result_code(self):
         if len(self.args) > 0:
-            error(pos, "Positive length tuples must be constructed.")
+            error(self.pos, "Positive length tuples must be constructed.")
         else:
             return Naming.empty_tuple
 
@@ -2875,6 +2853,9 @@ class SizeofVarNode(SizeofNode):
 #
 #-------------------------------------------------------------------
 
+def _not_in(x, seq):
+    return x not in seq
+
 compile_time_binary_operators = {
     '<': operator.lt,
     '<=': operator.le,
@@ -2897,8 +2878,8 @@ compile_time_binary_operators = {
     '-': operator.sub,
     #'/': operator.truediv,
     '^': operator.xor,
-    'in': lambda x, y: x in y,
-    'not_in': lambda x, y: x not in y,
+    'in': operator.contains,
+    'not_in': _not_in,
 }
 
 def get_compile_time_binop(node):

Cython/Compiler/Main.py

@@ -4,7 +4,7 @@
 
 import os, sys, re, codecs
 if sys.version_info[:2] < (2, 2):
-    print >>sys.stderr, "Sorry, Cython requires Python 2.2 or later"
+    sys.stderr.write("Sorry, Cython requires Python 2.2 or later\n")
     sys.exit(1)
 
 from time import time
@@ -146,7 +146,7 @@ class Context:
         else:
             filename_encoding = sys.getfilesystemencoding()
             if filename_encoding is None:
-                filename_encoding = getdefaultencoding()
+                filename_encoding = sys.getdefaultencoding()
             name = source_filename.decode(filename_encoding)
 
         s = PyrexScanner(f, name, source_encoding = f.encoding,
@@ -327,7 +327,7 @@ def main(command_line = 0):
         options = default_options
         sources = args
     if options.show_version:
-        print >>sys.stderr, "Cython version %s" % Version.version
+        sys.stderr.write("Cython version %s\n" % Version.version)
     if options.working_path!="":
         os.chdir(options.working_path)
     context = Context(options.include_path)
@@ -337,7 +337,7 @@ def main(command_line = 0):
             if result.num_errors > 0:
                 any_failures = 1
         except PyrexError, e:
-            print >>sys.stderr, e
+            sys.stderr.write(str(e) + '\n')
             any_failures = 1
     if any_failures:
         sys.exit(1)

Cython/Compiler/Nodes.py

@@ -12,7 +12,7 @@ import TypeSlots
 from PyrexTypes import py_object_type, error_type, CTypedefType, CFuncType
 from Symtab import ModuleScope, LocalScope, \
     StructOrUnionScope, PyClassScope, CClassScope
-from Cython.Utils import open_new_file, replace_suffix
+from Cython.Utils import open_new_file, replace_suffix, EncodedString
 import Options
 import ControlFlow
 
@@ -41,10 +41,10 @@ def relative_position(pos):
 def embed_position(pos, docstring):
     if not Options.embed_pos_in_docstring:
         return docstring
-    pos_line = u'File: %s (starting at line %s)' % relative_position(self.pos)
+    pos_line = u'File: %s (starting at line %s)' % relative_position(pos)
     if docstring is None:
         # unicode string
-        return ExprNodes.EncodedString(pos_line)
+        return EncodedString(pos_line)
 
     # make sure we can encode the filename in the docstring encoding
     # otherwise make the docstring a unicode string
@@ -57,13 +57,13 @@ def embed_position(pos, docstring):
 
     if not docstring:
         # reuse the string encoding of the original docstring
-        doc = ExprNodes.EncodedString(pos_line)
+        doc = EncodedString(pos_line)
     else:
-        doc = ExprNodes.EncodedString(pos_line + u'\\n' + docstring)
+        doc = EncodedString(pos_line + u'\\n' + docstring)
     doc.encoding = encoding
     return doc
 
-class AttributeAccessor:
+class _AttributeAccessor(object):
     """Used as the result of the Node.get_children_accessors() generator"""
     def __init__(self, obj, attrname):
         self.obj = obj
@@ -78,7 +78,18 @@ class AttributeAccessor:
     def name(self):
         return self.attrname
 
-class Node:
+class _AttributeIterator(object):
+    """Used as the result of the Node.get_children_accessors() generator"""
+    def __init__(self, obj, attrnames):
+        self.obj = obj
+        self.attrnames = iter(attrnames)
+    def __iter__(self):
+        return self
+    def __next__(self):
+        return _AttributeAccessor(self.obj, self.attrnames.next())
+    next = __next__
+
+class Node(object):
     #  pos         (string, int, int)   Source file position
     #  is_name     boolean              Is a NameNode
     #  is_literal  boolean              Is a ConstNode
@@ -129,13 +140,7 @@ class Node:
         if attrnames is None:
             raise InternalError("Children access not implemented for %s" % \
                 self.__class__.__name__)
-        for name in attrnames:
-            a = AttributeAccessor(self, name)
-            yield a
-            # Not really needed, but one wants to enforce clients not to
-            # hold on to iterated objects, in case more advanced iteration
-            # is needed later
-            a.attrname = None
+        return _AttributeIterator(self, attrnames)
     
     def get_child_attrs(self):
         """Utility method for more easily implementing get_child_accessors.

Cython/Compiler/Parsing.py

@@ -2,7 +2,7 @@
 #   Pyrex Parser
 #
 
-import os, re, codecs
+import os, re
 from string import join, replace
 from types import ListType, TupleType
 from Scanning import PyrexScanner
@@ -282,7 +282,7 @@ def p_call(s, function):
             if not arg.is_name:
                 s.error("Expected an identifier before '='",
                     pos = arg.pos)
-            encoded_name = ExprNodes.EncodedString(arg.name)
+            encoded_name = Utils.EncodedString(arg.name)
             encoded_name.encoding = s.source_encoding
             keyword = ExprNodes.StringNode(arg.pos, 
                 value = encoded_name)
@@ -503,11 +503,11 @@ def p_name(s, name):
             elif isinstance(value, float):
                 return ExprNodes.FloatNode(pos, value = rep)
             elif isinstance(value, str):
-                sval = ExprNodes.EncodedString(rep[1:-1])
+                sval = Utils.EncodedString(rep[1:-1])
                 sval.encoding = value.encoding
                 return ExprNodes.StringNode(pos, value = sval)
             elif isinstance(value, unicode):
-                sval = ExprNodes.EncodedString(rep[2:-1])
+                sval = Utils.EncodedString(rep[2:-1])
                 return ExprNodes.StringNode(pos, value = sval)
             else:
                 error(pos, "Invalid type for compile-time constant: %s"
@@ -523,12 +523,12 @@ def p_cat_string_literal(s):
         while s.sy == 'BEGIN_STRING':
             next_kind, next_value = p_string_literal(s)
             if next_kind == 'c':
-                self.error(
-                    "Cannot concatenate char literal with another string or char literal")
+                error(s.position(),
+                      "Cannot concatenate char literal with another string or char literal")
             elif next_kind == 'u':
                 kind = 'u'
             strings.append(next_value)
-        value = ExprNodes.EncodedString( u''.join(strings) )
+        value = Utils.EncodedString( u''.join(strings) )
         if kind != 'u':
             value.encoding = s.source_encoding
     return kind, value
@@ -600,7 +600,7 @@ def p_string_literal(s):
                 "Unexpected token %r:%r in string literal" %
                     (sy, s.systring))
     s.next()
-    value = ExprNodes.EncodedString( u''.join(chars) )
+    value = Utils.EncodedString( u''.join(chars) )
     if kind != 'u':
         value.encoding = s.source_encoding
     #print "p_string_literal: value =", repr(value) ###
@@ -915,7 +915,7 @@ def p_import_statement(s):
                     ExprNodes.StringNode(pos, value = "*")])
             else:
                 name_list = None
-            dotted_name = ExprNodes.EncodedString(dotted_name)
+            dotted_name = Utils.EncodedString(dotted_name)
             dotted_name.encoding = s.source_encoding
             stat = Nodes.SingleAssignmentNode(pos,
                 lhs = ExprNodes.NameNode(pos, 
@@ -955,7 +955,7 @@ def p_from_import_statement(s):
         imported_name_strings = []
         items = []
         for (name_pos, name, as_name) in imported_names:
-            encoded_name = ExprNodes.EncodedString(name)
+            encoded_name = Utils.EncodedString(name)
             encoded_name.encoding = s.source_encoding
             imported_name_strings.append(
                 ExprNodes.StringNode(name_pos, value = encoded_name))
@@ -965,7 +965,7 @@ def p_from_import_statement(s):
                  	name = as_name or name)))
         import_list = ExprNodes.ListNode(
             imported_names[0][0], args = imported_name_strings)
-        dotted_name = ExprNodes.EncodedString(dotted_name)
+        dotted_name = Utils.EncodedString(dotted_name)
         dotted_name.encoding = s.source_encoding
         return Nodes.FromImportStatNode(pos,
             module = ExprNodes.ImportNode(dotted_name_pos,
@@ -1971,7 +1971,7 @@ def p_class_statement(s):
     # s.sy == 'class'
     pos = s.position()
     s.next()
-    class_name = ExprNodes.EncodedString( p_ident(s) )
+    class_name = Utils.EncodedString( p_ident(s) )
     class_name.encoding = s.source_encoding
     if s.sy == '(':
         s.next()

Cython/Compiler/Scanning.py

@@ -59,6 +59,7 @@ def open_pickled_lexicon(expected_hash):
     # Try to open pickled lexicon file and verify that
     # it matches the source file. Returns the opened
     # file if successful, otherwise None. ???
+    global lexicon_pickle
     f = None
     result = None
     if os.path.exists(lexicon_pickle):

Cython/Compiler/Symtab.py

@@ -7,7 +7,7 @@ from Errors import warning, error, InternalError
 import Options
 import Naming
 import PyrexTypes
-from PyrexTypes import *
+from PyrexTypes import py_object_type
 import TypeSlots
 from TypeSlots import \
     pyfunction_signature, pymethod_signature, \
@@ -299,7 +299,8 @@ class Scope:
                 cname = self.mangle(Naming.type_prefix, name)
         entry = self.lookup_here(name)
         if not entry:
-            type = CStructOrUnionType(name, kind, scope, typedef_flag, cname)
+            type = PyrexTypes.CStructOrUnionType(
+                name, kind, scope, typedef_flag, cname)
             entry = self.declare_type(name, type, pos, cname,
                 visibility = visibility, defining = scope is not None)
             self.sue_entries.append(entry)
@@ -336,7 +337,7 @@ class Scope:
                     cname = name
                 else:
                     cname = self.mangle(Naming.type_prefix, name)
-            type = CEnumType(name, cname, typedef_flag)
+            type = PyrexTypes.CEnumType(name, cname, typedef_flag)
         else:
             type = PyrexTypes.c_anon_enum_type
         entry = self.declare_type(name, type, pos, cname = cname,
@@ -437,10 +438,10 @@ class Scope:
         # Add an entry for a string constant.
         cname = self.new_const_cname()
         if value.is_unicode:
-            c_type = c_utf8_char_array_type
+            c_type = PyrexTypes.c_utf8_char_array_type
             value = value.utf8encode()
         else:
-            c_type = c_char_array_type
+            c_type = PyrexTypes.c_char_array_type
             value = value.byteencode()
         entry = Entry("", cname, c_type, init = value)
         entry.used = 1
@@ -494,16 +495,6 @@ class Scope:
             genv.pynum_entries.append(entry)
         return entry
         
-    def add_py_obj(self, obj, c_prefix=''):
-        obj.check_const()
-        cname = self.new_const_cname(c_prefix)
-        entry = Entry("", cname, py_object_type, init = value)
-        entry.used = 1
-        entry.is_interned = 1
-        self.const_entries.append(entry)
-        self.interned_objs.append(entry)
-        return entry
-        
     def get_py_obj(self, obj, c_prefix=''):
         # Get entry for a generic constant. Returns an existing
         # one if possible, otherwise creates a new one.
@@ -513,7 +504,6 @@ class Scope:
             entry = genv.add_py_num(obj, c_prefix)
             genv.obj_to_entry[obj] = entry
         return entry
-        
     
     def new_const_cname(self):
         # Create a new globally-unique name for a constant.
@@ -532,7 +522,7 @@ class Scope:
         cname = "%s%d" % (Naming.pyrex_prefix, n)
         entry = Entry("", cname, type)
         entry.used = 1
-        if type.is_pyobject or type == c_py_ssize_t_type:
+        if type.is_pyobject or type == PyrexTypes.c_py_ssize_t_type:
             entry.init = "0"
         self.cname_to_entry[entry.cname] = entry
         self.temp_entries.append(entry)
@@ -844,8 +834,9 @@ class ModuleScope(Scope):
         self.default_entries.append(entry)
         return entry
         
-    def new_const_cname(self, prefix=''):
+    def new_const_cname(self):
         # Create a new globally-unique name for a constant.
+        prefix=''
         n = self.const_counter
         self.const_counter = n + 1
         return "%s%s_%d" % (Naming.const_prefix, prefix, n)
@@ -877,7 +868,7 @@ class ModuleScope(Scope):
         # Make a new entry if needed
         #
         if not entry:
-            type = PyExtensionType(name, typedef_flag, base_type)
+            type = PyrexTypes.PyExtensionType(name, typedef_flag, base_type)
             type.pos = pos
             if visibility == 'extern':
                 type.module_name = module_name
@@ -1075,7 +1066,7 @@ class StructOrUnionScope(Scope):
         if not cname:
             cname = name
         if type.is_cfunction:
-            type = CPtrType(type)
+            type = PyrexTypes.CPtrType(type)
         entry = self.declare(name, cname, type, pos)
         entry.is_variable = 1
         self.var_entries.append(entry)
@@ -1115,9 +1106,12 @@ class ClassScope(Scope):
             # Don't want to add a cfunction to this scope 'cause that would mess with 
             # the type definition, so we just return the right entry. 
             self.use_utility_code(classmethod_utility_code)
-            entry = Entry("classmethod", 
-                          "__Pyx_Method_ClassMethod", 
-                          CFuncType(py_object_type, [CFuncTypeArg("", py_object_type, None)], 0, 0))
+            entry = Entry(
+                "classmethod", 
+                "__Pyx_Method_ClassMethod", 
+                PyrexTypes.CFuncType(
+                    py_object_type,
+                    [PyrexTypes.CFuncTypeArg("", py_object_type, None)], 0, 0))
             entry.is_cfunction = 1
             return entry
         else:

Cython/Compiler/Transform.py

@@ -57,7 +57,7 @@ class Transform(object):
         parent has with child. This method should always return the node which the parent
         should use for this relation, which can either be the same node, None to remove
         the node, or a different node."""
-        raise InternalError("Not implemented")
+        raise NotImplementedError("Not implemented")
 
 class PrintTree(Transform):
     """Prints a representation of the tree to standard output.

Cython/Distutils/build_ext.py

@@ -85,9 +85,9 @@ class build_ext(_build_ext.build_ext):
 
         """
         Walk the list of source files in 'sources', looking for Cython
-        source (.pyx) files.  Run Cython on all that are found, and return
-        a modified 'sources' list with Cython source files replaced by the
-        generated C (or C++) files.
+        source files (.pyx and .py).  Run Cython on all that are
+        found, and return a modified 'sources' list with Cython source
+        files replaced by the generated C (or C++) files.
         """
 
         if PyrexError == None:
@@ -158,6 +158,9 @@ class build_ext(_build_ext.build_ext):
         newest_dependency = None
         for source in sources:
             (base, ext) = os.path.splitext(os.path.basename(source))
+            if ext == ".py":
+                # FIXME: we might want to special case this some more
+                ext = '.pyx'
             if ext == ".pyx":			  # Cython source file
                 output_dir = target_dir or os.path.dirname(source)
                 new_sources.append(os.path.join(output_dir, base + target_ext))

Cython/Plex/Machines.py

@@ -258,7 +258,8 @@ class FastMachine:
   def ranges_to_string(self, range_list):
     return string.join(map(self.range_to_string, range_list), ",")
   
-  def range_to_string(self, (c1, c2)):
+  def range_to_string(self, range_tuple):
+    (c1, c2) = range_tuple
     if c1 == c2:
       return repr(c1)
     else:

Cython/Plex/Regexps.py

@@ -122,7 +122,7 @@ class RE:
         beginning of a line. If nocase is true, upper and lower case
         letters should be treated as equivalent.
         """
-        raise exceptions.UnimplementedMethod("%s.build_machine not implemented" % 
+        raise NotImplementedError("%s.build_machine not implemented" % 
             self.__class__.__name__)
     
     def build_opt(self, m, initial_state, c):

Cython/Plex/Traditional.py

@@ -6,7 +6,7 @@
 #
 #=======================================================================
 
-from Regexps import *
+from Regexps import Alt, Seq, Rep, Rep1, Opt, Any, AnyBut, Bol, Eol, Char
 from Errors import PlexError
 
 class RegexpSyntaxError(PlexError):
@@ -104,7 +104,7 @@ class REParser:
           char_list.append(chr(a))
       else:
         char_list.append(c1)
-    chars = string.join(char_list, "")
+    chars = ''.join(char_list)
     if invert:
       return AnyBut(chars)
     else:

Cython/Unix/LinuxSystem.py

@@ -11,7 +11,7 @@ import os, sys
 from Cython.Utils import replace_suffix
 from Cython.Compiler.Errors import PyrexError
 
-version = "%s.%s" % sys.version[:2]
+version = "%s.%s" % sys.version_info[:2]
 py_include_dirs = [
     "%s/include/python%s" % (sys.prefix, version)
 ]

Cython/Utils.py

@@ -54,3 +54,25 @@ def detect_file_encoding(source_filename):
 def open_source_file(source_filename, mode="rU"):
     encoding = detect_file_encoding(source_filename)
     return codecs.open(source_filename, mode=mode, encoding=encoding)
+
+class EncodedString(unicode):
+    # unicode string subclass to keep track of the original encoding.
+    # 'encoding' is None for unicode strings and the source encoding
+    # otherwise
+    encoding = None
+
+    def byteencode(self):
+        assert self.encoding is not None
+        return self.encode(self.encoding)
+
+    def utf8encode(self):
+        assert self.encoding is None
+        return self.encode("UTF-8")
+
+    def is_unicode(self):
+        return self.encoding is None
+    is_unicode = property(is_unicode)
+
+#    def __eq__(self, other):
+#        return unicode.__eq__(self, other) and \
+#            getattr(other, 'encoding', '') == self.encoding

tests/run/pylistsubtype.pyx

+__doc__ = """
+    >>> l1 = Sub1([1,2,3])
+    >>> len(l1)
+    3
+
+    >>> l2 = Sub2([1,2,3])
+    >>> len(l2)
+    3
+
+    >>> isinstance(l1, list)
+    True
+    >>> isinstance(l2, list)
+    True
+    >>> isinstance(l1, Sub1)
+    True
+    >>> isinstance(l1, Sub2)
+    True
+    >>> isinstance(l2, Sub1)
+    False
+    >>> isinstance(l2, Sub2)
+    True
+"""
+
+cdef extern from *:
+    ctypedef class __builtin__.list [ object PyListObject ]:
+        pass
+
+cdef class Sub2(list):
+    cdef char character
+
+cdef class Sub1(Sub2):
+    cdef char character

tests/run/unicodeliteralsdefault.pyx

+## keep two lines free to make sure PEP 263 does not apply
+##
+
+##
+
+# This file is written in UTF-8, but it has no encoding declaration,
+# so it just defaults to UTF-8 (PEP 3120).
+
+__doc__ = r"""
+    >>> sa
+    'abc'
+    >>> ua
+    u'abc'
+    >>> b
+    u'123'
+    >>> c
+    u'S\xf8k ik'
+    >>> d
+    u'\xfc\xd6\xe4'
+    >>> e
+    u'\x03g\xf8\uf8d2S\xf8k ik'
+    >>> f
+    u'\xf8'
+    >>> add
+    u'S\xf8k ik\xfc\xd6\xe4abc'
+    >>> null
+    u'\x00'
+""" + """
+    >>> len(sa)
+    3
+    >>> len(ua)
+    3
+    >>> len(b)
+    3
+    >>> len(c)
+    6
+    >>> len(d)
+    3
+    >>> len(e)
+    10
+    >>> len(f)
+    1
+    >>> len(add)
+    12
+    >>> len(null)
+    1
+""" + u"""
+    >>> sa == 'abc'
+    True
+    >>> ua == u'abc'
+    True
+    >>> b == u'123'
+    True
+    >>> c == u'Søk ik'
+    True
+    >>> d == u'üÖä'
+    True
+    >>> e == u'\x03\x67\xf8\uf8d2Søk ik'     # unescaped by Cython
+    True
+    >>> e == u'\\x03\\x67\\xf8\\uf8d2Søk ik' # unescaped by Python
+    True
+    >>> f == u'\xf8'  # unescaped by Cython
+    True
+    >>> f == u'\\xf8' # unescaped by Python
+    True
+    >>> add == u'Søk ik' + u'üÖä' + 'abc'
+    True
+    >>> null == u'\\x00' # unescaped by Python (required by doctest)
+    True
+"""
+
+sa = 'abc'
+ua = u'abc'
+
+b = u'123'
+c = u'Søk ik'
+d = u'üÖä'
+e = u'\x03\x67\xf8\uf8d2Søk ik'
+f = u'\xf8'
+
+add = u'Søk ik' + u'üÖä' + 'abc'
+null = u'\x00'

tests/run/unicodeliteralslatin1.pyx

+# -*- coding: latin-1 -*-
+
+__doc__ = r"""
+    >>> sa
+    'abc'
+    >>> ua
+    u'abc'
+    >>> b
+    u'123'
+    >>> c
+    u'S\xf8k ik'
+    >>> d
+    u'\xfc\xd6\xe4'
+    >>> e
+    u'\x03g\xf8\uf8d2S\xf8k ik'
+    >>> f
+    u'\xf8'
+    >>> add
+    u'S\xf8k ik\xfc\xd6\xe4abc'
+    >>> null
+    u'\x00'
+""" + """
+    >>> len(sa)
+    3
+    >>> len(ua)
+    3
+    >>> len(b)
+    3
+    >>> len(c)
+    6
+    >>> len(d)
+    3
+    >>> len(e)
+    10
+    >>> len(f)
+    1
+    >>> len(add)
+    12
+    >>> len(null)
+    1
+""" + u"""
+    >>> sa == 'abc'
+    True
+    >>> ua == u'abc'
+    True
+    >>> b == u'123'
+    True
+    >>> c == u'Sk ik'
+    True
+    >>> d == u''
+    True
+    >>> e == u'\x03\x67\xf8\uf8d2Sk ik'     # unescaped by Cython
+    True
+    >>> e == u'\\x03\\x67\\xf8\\uf8d2Sk ik' # unescaped by Python
+    True
+    >>> f == u'\xf8'  # unescaped by Cython
+    True
+    >>> f == u'\\xf8' # unescaped by Python
+    True
+    >>> add == u'Sk ik' + u'' + 'abc'
+    True
+    >>> null == u'\\x00' # unescaped by Python (required by doctest)
+    True
+"""
+
+sa = 'abc'
+ua = u'abc'
+
+b = u'123'
+c = u'Sk ik'
+d = u''
+e = u'\x03\x67\xf8\uf8d2Sk ik'
+f = u'\xf8'
+
+add = u'Sk ik' + u'' + 'abc'
+null = u'\x00'