* Migrate set() and frozenset() from the sandbox.

* Install the unittests, docs, newsitem, include file, and makefile update.
* Exercise the new functions whereever sets.py was being used.

Includes the docs for libfuncs.tex.  Separate docs for the types are
forthcoming.

Doc/lib/libfuncs.tex

@@ -477,6 +477,17 @@ class C:
   and is known to vary.}
 \end{funcdesc}
 
+\begin{funcdesc}{frozenset}{\optional{iterable}}
+  Return a frozenset object whose elements are taken from \var{iterable}.
+  Frozensets are sets that have no update methods but can be hashed and
+  used as members of other sets or as dictionary keys.  The elements of
+  a frozenset must be immutable themselves.  To represent sets of sets,
+  the inner sets should also be \class{frozenset} objects.  If
+  \var{iterable} is not specified, returns a new empty set,
+  \code{frozenset([])}.
+  \versionadded{2.4}  
+\end{funcdesc}
+
 \begin{funcdesc}{getattr}{object, name\optional{, default}}
   Return the value of the named attributed of \var{object}.  \var{name}
   must be a string.  If the string is the name of one of the object's
@@ -897,6 +908,14 @@ class C(object):
   \code{round(0.5)} is \code{1.0} and \code{round(-0.5)} is \code{-1.0}).
 \end{funcdesc}
 
+\begin{funcdesc}{set}{\optional{iterable}}
+  Return a set whose elements are taken from \var{iterable}.  The elements
+  must be immutable.  To represent sets of sets, the inner sets should
+  be \class{frozenset} objects.  If \var{iterable} is not specified,
+  returns a new empty set, \code{set([])}.
+  \versionadded{2.4}  
+\end{funcdesc}
+
 \begin{funcdesc}{setattr}{object, name, value}
   This is the counterpart of \function{getattr()}.  The arguments are an
   object, a string and an arbitrary value.  The string may name an

Include/Python.h

@@ -86,6 +86,7 @@
 #include "listobject.h"
 #include "dictobject.h"
 #include "enumobject.h"
+#include "setobject.h"
 #include "methodobject.h"
 #include "moduleobject.h"
 #include "funcobject.h"

Include/setobject.h

@@ -0,0 +1,26 @@
+
+/* Set object interface */
+
+#ifndef Py_SETOBJECT_H
+#define Py_SETOBJECT_H
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+This data structure is shared by set and frozenset objects.
+*/
+
+typedef struct {
+	PyObject_HEAD
+	PyObject *data;
+	long hash;	/* only used by frozenset objects */
+} PySetObject;
+
+PyAPI_DATA(PyTypeObject) PySet_Type;
+PyAPI_DATA(PyTypeObject) PyFrozenSet_Type;
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* !Py_SETOBJECT_H */

Lib/_strptime.py

@@ -16,7 +16,6 @@ import calendar
 from re import compile as re_compile
 from re import IGNORECASE
 from datetime import date as datetime_date
-from sets import ImmutableSet as sets_ImmutableSet
 try:
     from thread import allocate_lock as _thread_allocate_lock
 except:
@@ -165,11 +164,11 @@ class LocaleTime(object):
             time.tzset()
         except AttributeError:
             pass
-        no_saving = sets_ImmutableSet(["utc", "gmt", time.tzname[0].lower()])
+        no_saving = frozenset(["utc", "gmt", time.tzname[0].lower()])
         if time.daylight:
-            has_saving = sets_ImmutableSet([time.tzname[1].lower()])
+            has_saving = frozenset([time.tzname[1].lower()])
         else:
-            has_saving = sets_ImmutableSet()
+            has_saving = frozenset()
         self.timezone = (no_saving, has_saving)
 
 

Lib/test/regrtest.py

@@ -75,7 +75,6 @@ import traceback
 import random
 import cStringIO
 import warnings
-from sets import Set
 
 # I see no other way to suppress these warnings;
 # putting them in test_grammar.py has no effect:
@@ -306,7 +305,7 @@ def main(tests=None, testdir=None, verbose=0, quiet=0, generate=0,
         e = _ExpectedSkips()
         plat = sys.platform
         if e.isvalid():
-            surprise = Set(skipped) - e.getexpected() - Set(resource_denieds)
+            surprise = set(skipped) - e.getexpected() - set(resource_denieds)
             if surprise:
                 print count(len(surprise), "skip"), \
                       "unexpected on", plat + ":"
@@ -948,7 +947,7 @@ class _ExpectedSkips:
         self.valid = False
         if sys.platform in _expectations:
             s = _expectations[sys.platform]
-            self.expected = Set(s.split())
+            self.expected = set(s.split())
 
             if not os.path.supports_unicode_filenames:
                 self.expected.add('test_pep277')

Lib/test/test___all__.py

@@ -2,7 +2,6 @@ import unittest
 from test import test_support
 
 from test.test_support import verify, verbose
-from sets import Set
 import sys
 import warnings
 
@@ -43,8 +42,8 @@ class AllTest(unittest.TestCase):
         exec "from %s import *" % modname in names
         if names.has_key("__builtins__"):
             del names["__builtins__"]
-        keys = Set(names)
+        keys = set(names)
-        all = Set(sys.modules[modname].__all__)
+        all = set(sys.modules[modname].__all__)
         verify(keys==all, "%s != %s" % (keys, all))
 
     def test_all(self):

Lib/test/test_builtin.py

@@ -2,7 +2,6 @@
 
 import test.test_support, unittest
 from test.test_support import fcmp, have_unicode, TESTFN, unlink
-from sets import Set
 
 import sys, warnings, cStringIO
 warnings.filterwarnings("ignore", "hex../oct.. of negative int",
@@ -1104,9 +1103,9 @@ class BuiltinTest(unittest.TestCase):
     get_vars_f2 = staticmethod(get_vars_f2)
 
     def test_vars(self):
-        self.assertEqual(Set(vars()), Set(dir()))
+        self.assertEqual(set(vars()), set(dir()))
         import sys
-        self.assertEqual(Set(vars(sys)), Set(dir(sys)))
+        self.assertEqual(set(vars(sys)), set(dir(sys)))
         self.assertEqual(self.get_vars_f0(), {})
         self.assertEqual(self.get_vars_f2(), {'a': 1, 'b': 2})
         self.assertRaises(TypeError, vars, 42, 42)

Lib/test/test_enumerate.py

@@ -1,5 +1,4 @@
 import unittest
-from sets import Set
 
 from test import test_support
 
@@ -105,8 +104,8 @@ class EnumerateTestCase(unittest.TestCase):
     def test_tuple_reuse(self):
         # Tests an implementation detail where tuple is reused
         # whenever nothing else holds a reference to it
-        self.assertEqual(len(Set(map(id, list(enumerate(self.seq))))), len(self.seq))
+        self.assertEqual(len(set(map(id, list(enumerate(self.seq))))), len(self.seq))
-        self.assertEqual(len(Set(map(id, enumerate(self.seq)))), min(1,len(self.seq)))
+        self.assertEqual(len(set(map(id, enumerate(self.seq)))), min(1,len(self.seq)))
 
 class MyEnum(enumerate):
     pass

Lib/test/test_glob.py

@@ -2,7 +2,6 @@ import unittest
 from test.test_support import run_unittest, TESTFN
 import glob
 import os
-from sets import Set
 
 def mkdirs(fname):
     if os.path.exists(fname) or fname == '':
@@ -62,7 +61,7 @@ class GlobTests(unittest.TestCase):
         return glob.glob(p)
 
     def assertSequencesEqual_noorder(self, l1, l2):
-        self.assertEqual(Set(l1), Set(l2))
+        self.assertEqual(set(l1), set(l2))
 
     def test_glob_literal(self):
         eq = self.assertSequencesEqual_noorder

Lib/test/test_pyclbr.py

@@ -7,7 +7,6 @@ import unittest, sys
 from types import ClassType, FunctionType, MethodType
 import pyclbr
 from unittest import TestCase
-from sets import Set
 
 # This next line triggers an error on old versions of pyclbr.
 
@@ -24,7 +23,7 @@ class PyclbrTest(TestCase):
 
     def assertListEq(self, l1, l2, ignore):
         ''' succeed iff {l1} - {ignore} == {l2} - {ignore} '''
-        missing = (Set(l1) ^ Set(l2)) - Set(ignore)
+        missing = (set(l1) ^ set(l2)) - set(ignore)
         if missing:
             print >>sys.stderr, "l1=%r\nl2=%r\nignore=%r" % (l1, l2, ignore)
             self.fail("%r missing" % missing.pop())

Lib/test/test_random.py

@@ -6,7 +6,6 @@ import time
 import pickle
 import warnings
 from math import log, exp, sqrt, pi
-from sets import Set
 from test import test_support
 
 class TestBasicOps(unittest.TestCase):
@@ -64,9 +63,9 @@ class TestBasicOps(unittest.TestCase):
         for k in xrange(N+1):
             s = self.gen.sample(population, k)
             self.assertEqual(len(s), k)
-            uniq = Set(s)
+            uniq = set(s)
             self.assertEqual(len(uniq), k)
-            self.failUnless(uniq <= Set(population))
+            self.failUnless(uniq <= set(population))
         self.assertEqual(self.gen.sample([], 0), [])  # test edge case N==k==0
 
     def test_sample_distribution(self):
@@ -89,8 +88,7 @@ class TestBasicOps(unittest.TestCase):
 
     def test_sample_inputs(self):
         # SF bug #801342 -- population can be any iterable defining __len__()
-        from sets import Set
+        self.gen.sample(set(range(20)), 2)
-        self.gen.sample(Set(range(20)), 2)
         self.gen.sample(range(20), 2)
         self.gen.sample(xrange(20), 2)
         self.gen.sample(dict.fromkeys('abcdefghijklmnopqrst'), 2)
@@ -256,8 +254,8 @@ class MersenneTwister_TestBasicOps(TestBasicOps):
 
     def test_rangelimits(self):
         for start, stop in [(-2,0), (-(2**60)-2,-(2**60)), (2**60,2**60+2)]:
-            self.assertEqual(Set(range(start,stop)),
+            self.assertEqual(set(range(start,stop)),
-                Set([self.gen.randrange(start,stop) for i in xrange(100)]))
+                set([self.gen.randrange(start,stop) for i in xrange(100)]))
 
     def test_genrandbits(self):
         # Verify cross-platform repeatability
@@ -364,7 +362,7 @@ class TestModule(unittest.TestCase):
 
     def test__all__(self):
         # tests validity but not completeness of the __all__ list
-        self.failUnless(Set(random.__all__) <= Set(dir(random)))
+        self.failUnless(set(random.__all__) <= set(dir(random)))
 
 def test_main(verbose=None):
     testclasses =    (WichmannHill_TestBasicOps,

Lib/test/test_sort.py

@@ -1,7 +1,6 @@
 from test.test_support import verbose
 import random
 from UserList import UserList
-from sets import Set
 
 nerrors = 0
 
@@ -230,7 +229,7 @@ class TestSorted(unittest.TestCase):
             self.assertEqual(list.sorted(s), list.sorted(T(s)))
 
         s = ''.join(dict.fromkeys(s).keys())  # unique letters only
-        for T in [unicode, Set, list, tuple, dict.fromkeys]:
+        for T in [unicode, set, frozenset, list, tuple, dict.fromkeys]:
             self.assertEqual(list.sorted(s), list.sorted(T(s)))
 
     def test_baddecorator(self):

Lib/test/test_strptime.py

@@ -4,7 +4,6 @@ import unittest
 import time
 import locale
 import re
-import sets
 import sys
 from test import test_support
 
@@ -167,7 +166,7 @@ class TimeRETests(unittest.TestCase):
         # Make sure when tuple or something has no values no regex is generated.
         # Fixes bug #661354
         test_locale = _strptime.LocaleTime()
-        test_locale.timezone = (sets.ImmutableSet(), sets.ImmutableSet())
+        test_locale.timezone = (frozenset(), frozenset())
         self.failUnless(_strptime.TimeRE(test_locale).pattern("%Z") == '',
                         "with timezone == ('',''), TimeRE().pattern('%Z') != ''")
 

Lib/test/test_userdict.py

@@ -1,8 +1,6 @@
 # Check every path through every method of UserDict
 
 import test.test_support, unittest
-from sets import Set
-
 import UserDict
 
 class TestMappingProtocol(unittest.TestCase):
@@ -69,7 +67,7 @@ class TestMappingProtocol(unittest.TestCase):
             self.assert_(hasattr(iter, 'next'))
             self.assert_(hasattr(iter, '__iter__'))
             x = list(iter)
-            self.assert_(Set(x)==Set(lst)==Set(ref))
+            self.assert_(set(x)==set(lst)==set(ref))
         check_iterandlist(d.iterkeys(), d.keys(), self.reference.keys())
         check_iterandlist(iter(d), d.keys(), self.reference.keys())
         check_iterandlist(d.itervalues(), d.values(), self.reference.values())
@@ -242,7 +240,7 @@ class UserDictTest(TestMappingProtocol):
         for k in u2:
             ikeys.append(k)
         keys = u2.keys()
-        self.assertEqual(Set(ikeys), Set(keys))
+        self.assertEqual(set(ikeys), set(keys))
 
         # Test setdefault
         t = UserDict.UserDict()

Lib/test/test_weakref.py

@@ -4,7 +4,6 @@ import UserList
 import weakref
 
 from test import test_support
-from sets import Set
 
 
 class C:
@@ -392,7 +391,7 @@ class MappingTestCase(TestBase):
                          "wrong object returned by weak dict!")
         items1 = dict.items()
         items2 = dict.copy().items()
-        self.assert_(Set(items1) == Set(items2),
+        self.assert_(set(items1) == set(items2),
                      "cloning of weak-keyed dictionary did not work!")
         del items1, items2
         self.assert_(len(dict) == self.COUNT)

Makefile.pre.in

@@ -275,6 +275,7 @@ OBJECT_OBJS=	\
 		Objects/object.o \
 		Objects/obmalloc.o \
 		Objects/rangeobject.o \
+                Objects/setobject.o \
 		Objects/sliceobject.o \
 		Objects/stringobject.o \
 		Objects/structseq.o \
@@ -500,6 +501,7 @@ PYTHON_HEADERS= \
 		Include/pystate.h \
 		Include/pythonrun.h \
 		Include/rangeobject.h \
+                Include/setobject.h \
 		Include/sliceobject.h \
 		Include/stringobject.h \
 		Include/structseq.h \

Misc/NEWS

@@ -12,6 +12,8 @@ What's New in Python 2.4 alpha 1?
 Core and builtins
 -----------------
 
+- Added two builtin types, set() and frozenset().
+
 - Critical bugfix, for SF bug 840829:  if cyclic garbage collection
   happened to occur during a weakref callback for a new-style class
   instance, subtle memory corruption was the result (in a release build;

PCbuild/pythoncore.dsp

@@ -467,6 +467,10 @@ SOURCE=..\Modules\rotormodule.c
 # End Source File
 # Begin Source File
 
+SOURCE=..\Objects\setobject.c
+# End Source File
+# Begin Source File
+
 SOURCE=..\Modules\shamodule.c
 # End Source File
 # Begin Source File

Python/bltinmodule.c

@@ -2116,12 +2116,14 @@ _PyBuiltin_Init(void)
 	SETBUILTIN("dict",		&PyDict_Type);
  	SETBUILTIN("enumerate",		&PyEnum_Type);
 	SETBUILTIN("float",		&PyFloat_Type);
+	SETBUILTIN("frozenset",		&PyFrozenSet_Type);
 	SETBUILTIN("property",		&PyProperty_Type);
 	SETBUILTIN("int",		&PyInt_Type);
 	SETBUILTIN("list",		&PyList_Type);
 	SETBUILTIN("long",		&PyLong_Type);
 	SETBUILTIN("object",		&PyBaseObject_Type);
 	SETBUILTIN("reversed",		&PyReversed_Type);
+	SETBUILTIN("set",		&PySet_Type);
 	SETBUILTIN("slice",		&PySlice_Type);
 	SETBUILTIN("staticmethod",	&PyStaticMethod_Type);
 	SETBUILTIN("str",		&PyString_Type);