diff --git a/tests/modeltests/basic/models.py b/tests/modeltests/basic/models.py
index ffce4cba6b..5f51fd8e11 100644
--- a/tests/modeltests/basic/models.py
+++ b/tests/modeltests/basic/models.py
@@ -15,411 +15,3 @@ class Article(models.Model):
 
     def __unicode__(self):
         return self.headline
-
-__test__ = {'API_TESTS': """
-# No articles are in the system yet.
->>> Article.objects.all()
-[]
-
-# Create an Article.
->>> from datetime import datetime
->>> a = Article(id=None, headline='Area man programs in Python', pub_date=datetime(2005, 7, 28))
-
-# Save it into the database. You have to call save() explicitly.
->>> a.save()
-
-# Now it has an ID. Note it's a long integer, as designated by the trailing "L".
->>> a.id
-1L
-
-# Models have a pk property that is an alias for the primary key attribute (by
-# default, the 'id' attribute).
->>> a.pk
-1L
-
-# Access database columns via Python attributes.
->>> a.headline
-'Area man programs in Python'
->>> a.pub_date
-datetime.datetime(2005, 7, 28, 0, 0)
-
-# Change values by changing the attributes, then calling save().
->>> a.headline = 'Area woman programs in Python'
->>> a.save()
-
-# Article.objects.all() returns all the articles in the database.
->>> Article.objects.all()
-[<Article: Area woman programs in Python>]
-
-# Django provides a rich database lookup API.
->>> Article.objects.get(id__exact=1)
-<Article: Area woman programs in Python>
->>> Article.objects.get(headline__startswith='Area woman')
-<Article: Area woman programs in Python>
->>> Article.objects.get(pub_date__year=2005)
-<Article: Area woman programs in Python>
->>> Article.objects.get(pub_date__year=2005, pub_date__month=7)
-<Article: Area woman programs in Python>
->>> Article.objects.get(pub_date__year=2005, pub_date__month=7, pub_date__day=28)
-<Article: Area woman programs in Python>
->>> Article.objects.get(pub_date__week_day=5)
-<Article: Area woman programs in Python>
-
-# The "__exact" lookup type can be omitted, as a shortcut.
->>> Article.objects.get(id=1)
-<Article: Area woman programs in Python>
->>> Article.objects.get(headline='Area woman programs in Python')
-<Article: Area woman programs in Python>
-
->>> Article.objects.filter(pub_date__year=2005)
-[<Article: Area woman programs in Python>]
->>> Article.objects.filter(pub_date__year=2004)
-[]
->>> Article.objects.filter(pub_date__year=2005, pub_date__month=7)
-[<Article: Area woman programs in Python>]
-
->>> Article.objects.filter(pub_date__week_day=5)
-[<Article: Area woman programs in Python>]
->>> Article.objects.filter(pub_date__week_day=6)
-[]
-
-# Django raises an Article.DoesNotExist exception for get() if the parameters
-# don't match any object.
->>> Article.objects.get(id__exact=2)
-Traceback (most recent call last):
-    ...
-DoesNotExist: Article matching query does not exist.
-
->>> Article.objects.get(pub_date__year=2005, pub_date__month=8)
-Traceback (most recent call last):
-    ...
-DoesNotExist: Article matching query does not exist.
-
->>> Article.objects.get(pub_date__week_day=6)
-Traceback (most recent call last):
-    ...
-DoesNotExist: Article matching query does not exist.
-
-# Lookup by a primary key is the most common case, so Django provides a
-# shortcut for primary-key exact lookups.
-# The following is identical to articles.get(id=1).
->>> Article.objects.get(pk=1)
-<Article: Area woman programs in Python>
-
-# pk can be used as a shortcut for the primary key name in any query
->>> Article.objects.filter(pk__in=[1])
-[<Article: Area woman programs in Python>]
-
-# Model instances of the same type and same ID are considered equal.
->>> a = Article.objects.get(pk=1)
->>> b = Article.objects.get(pk=1)
->>> a == b
-True
-
-# You can initialize a model instance using positional arguments, which should
-# match the field order as defined in the model.
->>> a2 = Article(None, 'Second article', datetime(2005, 7, 29))
->>> a2.save()
->>> a2.id
-2L
->>> a2.headline
-'Second article'
->>> a2.pub_date
-datetime.datetime(2005, 7, 29, 0, 0)
-
-# ...or, you can use keyword arguments.
->>> a3 = Article(id=None, headline='Third article', pub_date=datetime(2005, 7, 30))
->>> a3.save()
->>> a3.id
-3L
->>> a3.headline
-'Third article'
->>> a3.pub_date
-datetime.datetime(2005, 7, 30, 0, 0)
-
-# You can also mix and match position and keyword arguments, but be sure not to
-# duplicate field information.
->>> a4 = Article(None, 'Fourth article', pub_date=datetime(2005, 7, 31))
->>> a4.save()
->>> a4.headline
-'Fourth article'
-
-# Don't use invalid keyword arguments.
->>> a5 = Article(id=None, headline='Invalid', pub_date=datetime(2005, 7, 31), foo='bar')
-Traceback (most recent call last):
-    ...
-TypeError: 'foo' is an invalid keyword argument for this function
-
-# You can leave off the value for an AutoField when creating an object, because
-# it'll get filled in automatically when you save().
->>> a5 = Article(headline='Article 6', pub_date=datetime(2005, 7, 31))
->>> a5.save()
->>> a5.id
-5L
->>> a5.headline
-'Article 6'
-
-# If you leave off a field with "default" set, Django will use the default.
->>> a6 = Article(pub_date=datetime(2005, 7, 31))
->>> a6.save()
->>> a6.headline
-u'Default headline'
-
-# For DateTimeFields, Django saves as much precision (in seconds) as you
-# give it.
->>> a7 = Article(headline='Article 7', pub_date=datetime(2005, 7, 31, 12, 30))
->>> a7.save()
->>> Article.objects.get(id__exact=7).pub_date
-datetime.datetime(2005, 7, 31, 12, 30)
-
->>> a8 = Article(headline='Article 8', pub_date=datetime(2005, 7, 31, 12, 30, 45))
->>> a8.save()
->>> Article.objects.get(id__exact=8).pub_date
-datetime.datetime(2005, 7, 31, 12, 30, 45)
->>> a8.id
-8L
-
-# Saving an object again doesn't create a new object -- it just saves the old one.
->>> a8.save()
->>> a8.id
-8L
->>> a8.headline = 'Updated article 8'
->>> a8.save()
->>> a8.id
-8L
-
->>> a7 == a8
-False
->>> a8 == Article.objects.get(id__exact=8)
-True
->>> a7 != a8
-True
->>> Article.objects.get(id__exact=8) != Article.objects.get(id__exact=7)
-True
->>> Article.objects.get(id__exact=8) == Article.objects.get(id__exact=7)
-False
-
-# You can use 'in' to test for membership...
->>> a8 in Article.objects.all()
-True
-
-# ... but there will often be more efficient ways if that is all you need:
->>> Article.objects.filter(id=a8.id).exists()
-True
-
-# dates() returns a list of available dates of the given scope for the given field.
->>> Article.objects.dates('pub_date', 'year')
-[datetime.datetime(2005, 1, 1, 0, 0)]
->>> Article.objects.dates('pub_date', 'month')
-[datetime.datetime(2005, 7, 1, 0, 0)]
->>> Article.objects.dates('pub_date', 'day')
-[datetime.datetime(2005, 7, 28, 0, 0), datetime.datetime(2005, 7, 29, 0, 0), datetime.datetime(2005, 7, 30, 0, 0), datetime.datetime(2005, 7, 31, 0, 0)]
->>> Article.objects.dates('pub_date', 'day', order='ASC')
-[datetime.datetime(2005, 7, 28, 0, 0), datetime.datetime(2005, 7, 29, 0, 0), datetime.datetime(2005, 7, 30, 0, 0), datetime.datetime(2005, 7, 31, 0, 0)]
->>> Article.objects.dates('pub_date', 'day', order='DESC')
-[datetime.datetime(2005, 7, 31, 0, 0), datetime.datetime(2005, 7, 30, 0, 0), datetime.datetime(2005, 7, 29, 0, 0), datetime.datetime(2005, 7, 28, 0, 0)]
-
-# dates() requires valid arguments.
-
->>> Article.objects.dates()
-Traceback (most recent call last):
-   ...
-TypeError: dates() takes at least 3 arguments (1 given)
-
->>> Article.objects.dates('invalid_field', 'year')
-Traceback (most recent call last):
-   ...
-FieldDoesNotExist: Article has no field named 'invalid_field'
-
->>> Article.objects.dates('pub_date', 'bad_kind')
-Traceback (most recent call last):
-   ...
-AssertionError: 'kind' must be one of 'year', 'month' or 'day'.
-
->>> Article.objects.dates('pub_date', 'year', order='bad order')
-Traceback (most recent call last):
-   ...
-AssertionError: 'order' must be either 'ASC' or 'DESC'.
-
-# Use iterator() with dates() to return a generator that lazily requests each
-# result one at a time, to save memory.
->>> for a in Article.objects.dates('pub_date', 'day', order='DESC').iterator():
-...     print repr(a)
-datetime.datetime(2005, 7, 31, 0, 0)
-datetime.datetime(2005, 7, 30, 0, 0)
-datetime.datetime(2005, 7, 29, 0, 0)
-datetime.datetime(2005, 7, 28, 0, 0)
-
-# You can combine queries with & and |.
->>> s1 = Article.objects.filter(id__exact=1)
->>> s2 = Article.objects.filter(id__exact=2)
->>> s1 | s2
-[<Article: Area woman programs in Python>, <Article: Second article>]
->>> s1 & s2
-[]
-
-# You can get the number of objects like this:
->>> len(Article.objects.filter(id__exact=1))
-1
-
-# You can get items using index and slice notation.
->>> Article.objects.all()[0]
-<Article: Area woman programs in Python>
->>> Article.objects.all()[1:3]
-[<Article: Second article>, <Article: Third article>]
->>> s3 = Article.objects.filter(id__exact=3)
->>> (s1 | s2 | s3)[::2]
-[<Article: Area woman programs in Python>, <Article: Third article>]
-
-# Slicing works with longs.
->>> Article.objects.all()[0L]
-<Article: Area woman programs in Python>
->>> Article.objects.all()[1L:3L]
-[<Article: Second article>, <Article: Third article>]
->>> s3 = Article.objects.filter(id__exact=3)
->>> (s1 | s2 | s3)[::2L]
-[<Article: Area woman programs in Python>, <Article: Third article>]
-
-# And can be mixed with ints.
->>> Article.objects.all()[1:3L]
-[<Article: Second article>, <Article: Third article>]
-
-# Slices (without step) are lazy:
->>> Article.objects.all()[0:5].filter()
-[<Article: Area woman programs in Python>, <Article: Second article>, <Article: Third article>, <Article: Article 6>, <Article: Default headline>]
-
-# Slicing again works:
->>> Article.objects.all()[0:5][0:2]
-[<Article: Area woman programs in Python>, <Article: Second article>]
->>> Article.objects.all()[0:5][:2]
-[<Article: Area woman programs in Python>, <Article: Second article>]
->>> Article.objects.all()[0:5][4:]
-[<Article: Default headline>]
->>> Article.objects.all()[0:5][5:]
-[]
-
-# Some more tests!
->>> Article.objects.all()[2:][0:2]
-[<Article: Third article>, <Article: Article 6>]
->>> Article.objects.all()[2:][:2]
-[<Article: Third article>, <Article: Article 6>]
->>> Article.objects.all()[2:][2:3]
-[<Article: Default headline>]
-
-# Using an offset without a limit is also possible.
->>> Article.objects.all()[5:]
-[<Article: Fourth article>, <Article: Article 7>, <Article: Updated article 8>]
-
-# Also, once you have sliced you can't filter, re-order or combine
->>> Article.objects.all()[0:5].filter(id=1)
-Traceback (most recent call last):
-    ...
-AssertionError: Cannot filter a query once a slice has been taken.
-
->>> Article.objects.all()[0:5].order_by('id')
-Traceback (most recent call last):
-    ...
-AssertionError: Cannot reorder a query once a slice has been taken.
-
->>> Article.objects.all()[0:1] & Article.objects.all()[4:5]
-Traceback (most recent call last):
-    ...
-AssertionError: Cannot combine queries once a slice has been taken.
-
-# Negative slices are not supported, due to database constraints.
-# (hint: inverting your ordering might do what you need).
->>> Article.objects.all()[-1]
-Traceback (most recent call last):
-    ...
-AssertionError: Negative indexing is not supported.
->>> Article.objects.all()[0:-5]
-Traceback (most recent call last):
-    ...
-AssertionError: Negative indexing is not supported.
-
-# An Article instance doesn't have access to the "objects" attribute.
-# That's only available on the class.
->>> a7.objects.all()
-Traceback (most recent call last):
-    ...
-AttributeError: Manager isn't accessible via Article instances
-
->>> a7.objects
-Traceback (most recent call last):
-    ...
-AttributeError: Manager isn't accessible via Article instances
-
-# Bulk delete test: How many objects before and after the delete?
->>> Article.objects.all()
-[<Article: Area woman programs in Python>, <Article: Second article>, <Article: Third article>, <Article: Article 6>, <Article: Default headline>, <Article: Fourth article>, <Article: Article 7>, <Article: Updated article 8>]
->>> Article.objects.filter(id__lte=4).delete()
->>> Article.objects.all()
-[<Article: Article 6>, <Article: Default headline>, <Article: Article 7>, <Article: Updated article 8>]
-"""}
-
-from django.conf import settings
-
-if connection.features.supports_microsecond_precision:
-    __test__['API_TESTS'] += """
-# In PostgreSQL, microsecond-level precision is available.
->>> a9 = Article(headline='Article 9', pub_date=datetime(2005, 7, 31, 12, 30, 45, 180))
->>> a9.save()
->>> Article.objects.get(id__exact=9).pub_date
-datetime.datetime(2005, 7, 31, 12, 30, 45, 180)
-"""
-else:
-    __test__['API_TESTS'] += """
-# In MySQL, microsecond-level precision isn't available. You'll lose
-# microsecond-level precision once the data is saved.
->>> a9 = Article(headline='Article 9', pub_date=datetime(2005, 7, 31, 12, 30, 45, 180))
->>> a9.save()
->>> Article.objects.get(id__exact=9).pub_date
-datetime.datetime(2005, 7, 31, 12, 30, 45)
-"""
-
-__test__['API_TESTS'] += """
-
-# You can manually specify the primary key when creating a new object.
->>> a101 = Article(id=101, headline='Article 101', pub_date=datetime(2005, 7, 31, 12, 30, 45))
->>> a101.save()
->>> a101 = Article.objects.get(pk=101)
->>> a101.headline
-u'Article 101'
-
-# You can create saved objects in a single step
->>> a10 = Article.objects.create(headline="Article 10", pub_date=datetime(2005, 7, 31, 12, 30, 45))
->>> Article.objects.get(headline="Article 10")
-<Article: Article 10>
-
-# Edge-case test: A year lookup should retrieve all objects in the given
-year, including Jan. 1 and Dec. 31.
->>> a11 = Article.objects.create(headline='Article 11', pub_date=datetime(2008, 1, 1))
->>> a12 = Article.objects.create(headline='Article 12', pub_date=datetime(2008, 12, 31, 23, 59, 59, 999999))
->>> Article.objects.filter(pub_date__year=2008)
-[<Article: Article 11>, <Article: Article 12>]
-
-# Unicode data works, too.
->>> a = Article(headline=u'\u6797\u539f \u3081\u3050\u307f', pub_date=datetime(2005, 7, 28))
->>> a.save()
->>> Article.objects.get(pk=a.id).headline
-u'\u6797\u539f \u3081\u3050\u307f'
-
-# Model instances have a hash function, so they can be used in sets or as
-# dictionary keys. Two models compare as equal if their primary keys are equal.
->>> s = set([a10, a11, a12])
->>> Article.objects.get(headline='Article 11') in s
-True
-
-# The 'select' argument to extra() supports names with dashes in them, as long
-# as you use values().
->>> dicts = Article.objects.filter(pub_date__year=2008).extra(select={'dashed-value': '1'}).values('headline', 'dashed-value')
->>> [sorted(d.items()) for d in dicts]
-[[('dashed-value', 1), ('headline', u'Article 11')], [('dashed-value', 1), ('headline', u'Article 12')]]
-
-# If you use 'select' with extra() and names containing dashes on a query
-# that's *not* a values() query, those extra 'select' values will silently be
-# ignored.
->>> articles = Article.objects.filter(pub_date__year=2008).extra(select={'dashed-value': '1', 'undashedvalue': '2'})
->>> articles[0].undashedvalue
-2
-"""
diff --git a/tests/modeltests/basic/tests.py b/tests/modeltests/basic/tests.py
new file mode 100644
index 0000000000..966798d78b
--- /dev/null
+++ b/tests/modeltests/basic/tests.py
@@ -0,0 +1,554 @@
+from datetime import datetime
+
+from django.core.exceptions import ObjectDoesNotExist
+from django.db import models, DEFAULT_DB_ALIAS, connection
+from django.db.models.fields import FieldDoesNotExist
+from django.test import TestCase, skipIfDBFeature, skipUnlessDBFeature
+
+from models import Article
+
+
+class ModelTest(TestCase):
+
+    def test_lookup(self):
+        # No articles are in the system yet.
+        self.assertQuerysetEqual(Article.objects.all(), [])
+
+        # Create an Article.
+        a = Article(
+            id=None,
+            headline='Area man programs in Python',
+            pub_date=datetime(2005, 7, 28),
+        )
+
+        # Save it into the database. You have to call save() explicitly.
+        a.save()
+
+        # Now it has an ID.
+        self.assertTrue(a.id != None)
+
+        # Models have a pk property that is an alias for the primary key
+        # attribute (by default, the 'id' attribute).
+        self.assertEqual(a.pk, a.id)
+
+        # Access database columns via Python attributes.
+        self.assertEqual(a.headline, 'Area man programs in Python')
+        self.assertEqual(a.pub_date, datetime(2005, 7, 28, 0, 0))
+
+        # Change values by changing the attributes, then calling save().
+        a.headline = 'Area woman programs in Python'
+        a.save()
+
+        # Article.objects.all() returns all the articles in the database.
+        self.assertQuerysetEqual(Article.objects.all(),
+            ['<Article: Area woman programs in Python>'])
+
+        # Django provides a rich database lookup API.
+        self.assertEqual(Article.objects.get(id__exact=a.id), a)
+        self.assertEqual(Article.objects.get(headline__startswith='Area woman'), a)
+        self.assertEqual(Article.objects.get(pub_date__year=2005), a)
+        self.assertEqual(Article.objects.get(pub_date__year=2005, pub_date__month=7), a)
+        self.assertEqual(Article.objects.get(pub_date__year=2005, pub_date__month=7, pub_date__day=28), a)
+        self.assertEqual(Article.objects.get(pub_date__week_day=5), a)
+
+        # The "__exact" lookup type can be omitted, as a shortcut.
+        self.assertEqual(Article.objects.get(id=a.id), a)
+        self.assertEqual(Article.objects.get(headline='Area woman programs in Python'), a)
+
+        self.assertQuerysetEqual(
+            Article.objects.filter(pub_date__year=2005),
+            ['<Article: Area woman programs in Python>'],
+        )
+        self.assertQuerysetEqual(
+            Article.objects.filter(pub_date__year=2004),
+            [],
+        )
+        self.assertQuerysetEqual(
+            Article.objects.filter(pub_date__year=2005, pub_date__month=7),
+            ['<Article: Area woman programs in Python>'],
+        )
+
+        self.assertQuerysetEqual(
+            Article.objects.filter(pub_date__week_day=5),
+            ['<Article: Area woman programs in Python>'],
+        )
+        self.assertQuerysetEqual(
+            Article.objects.filter(pub_date__week_day=6),
+            [],
+        )
+
+        # Django raises an Article.DoesNotExist exception for get() if the
+        # parameters don't match any object.
+        self.assertRaisesRegexp(
+            ObjectDoesNotExist,
+            "Article matching query does not exist.",
+            Article.objects.get,
+            id__exact=2000,
+        )
+
+        self.assertRaisesRegexp(
+            ObjectDoesNotExist,
+            "Article matching query does not exist.",
+            Article.objects.get,
+            pub_date__year=2005,
+            pub_date__month=8,
+        )
+
+        self.assertRaisesRegexp(
+            ObjectDoesNotExist,
+            "Article matching query does not exist.",
+            Article.objects.get,
+            pub_date__week_day=6,
+        )
+
+        # Lookup by a primary key is the most common case, so Django
+        # provides a shortcut for primary-key exact lookups.
+        # The following is identical to articles.get(id=a.id).
+        self.assertEqual(Article.objects.get(pk=a.id), a)
+
+        # pk can be used as a shortcut for the primary key name in any query.
+        self.assertQuerysetEqual(Article.objects.filter(pk__in=[a.id]),
+            ["<Article: Area woman programs in Python>"])
+
+        # Model instances of the same type and same ID are considered equal.
+        a = Article.objects.get(pk=a.id)
+        b = Article.objects.get(pk=a.id)
+        self.assertEqual(a, b)
+
+    def test_object_creation(self):
+        # Create an Article.
+        a = Article(
+            id=None,
+            headline='Area man programs in Python',
+            pub_date=datetime(2005, 7, 28),
+        )
+
+        # Save it into the database. You have to call save() explicitly.
+        a.save()
+
+        # You can initialize a model instance using positional arguments,
+        # which should match the field order as defined in the model.
+        a2 = Article(None, 'Second article', datetime(2005, 7, 29))
+        a2.save()
+
+        self.assertNotEqual(a2.id, a.id)
+        self.assertEqual(a2.headline, 'Second article')
+        self.assertEqual(a2.pub_date, datetime(2005, 7, 29, 0, 0))
+
+        # ...or, you can use keyword arguments.
+        a3 = Article(
+            id=None,
+            headline='Third article',
+            pub_date=datetime(2005, 7, 30),
+        )
+        a3.save()
+
+        self.assertNotEqual(a3.id, a.id)
+        self.assertNotEqual(a3.id, a2.id)
+        self.assertEqual(a3.headline, 'Third article')
+        self.assertEqual(a3.pub_date, datetime(2005, 7, 30, 0, 0))
+
+        # You can also mix and match position and keyword arguments, but
+        # be sure not to duplicate field information.
+        a4 = Article(None, 'Fourth article', pub_date=datetime(2005, 7, 31))
+        a4.save()
+        self.assertEqual(a4.headline, 'Fourth article')
+
+        # Don't use invalid keyword arguments.
+        self.assertRaisesRegexp(
+            TypeError,
+            "'foo' is an invalid keyword argument for this function",
+            Article,
+            id=None,
+            headline='Invalid',
+            pub_date=datetime(2005, 7, 31),
+            foo='bar',
+        )
+
+        # You can leave off the value for an AutoField when creating an
+        # object, because it'll get filled in automatically when you save().
+        a5 = Article(headline='Article 6', pub_date=datetime(2005, 7, 31))
+        a5.save()
+        self.assertEqual(a5.headline, 'Article 6')
+
+        # If you leave off a field with "default" set, Django will use
+        # the default.
+        a6 = Article(pub_date=datetime(2005, 7, 31))
+        a6.save()
+        self.assertEqual(a6.headline, u'Default headline')
+
+        # For DateTimeFields, Django saves as much precision (in seconds)
+        # as you give it.
+        a7 = Article(
+            headline='Article 7',
+            pub_date=datetime(2005, 7, 31, 12, 30),
+        )
+        a7.save()
+        self.assertEqual(Article.objects.get(id__exact=a7.id).pub_date,
+            datetime(2005, 7, 31, 12, 30))
+
+        a8 = Article(
+            headline='Article 8',
+            pub_date=datetime(2005, 7, 31, 12, 30, 45),
+        )
+        a8.save()
+        self.assertEqual(Article.objects.get(id__exact=a8.id).pub_date,
+            datetime(2005, 7, 31, 12, 30, 45))
+
+        # Saving an object again doesn't create a new object -- it just saves
+        # the old one.
+        current_id = a8.id
+        a8.save()
+        self.assertEqual(a8.id, current_id)
+        a8.headline = 'Updated article 8'
+        a8.save()
+        self.assertEqual(a8.id, current_id)
+
+        # Check that != and == operators behave as expecte on instances
+        self.assertTrue(a7 != a8)
+        self.assertFalse(a7 == a8)
+        self.assertEqual(a8, Article.objects.get(id__exact=a8.id))
+
+        self.assertTrue(Article.objects.get(id__exact=a8.id) != Article.objects.get(id__exact=a7.id))
+        self.assertFalse(Article.objects.get(id__exact=a8.id) == Article.objects.get(id__exact=a7.id))
+
+        # You can use 'in' to test for membership...
+        self.assertTrue(a8 in Article.objects.all())
+
+        # ... but there will often be more efficient ways if that is all you need:
+        self.assertTrue(Article.objects.filter(id=a8.id).exists())
+
+        # dates() returns a list of available dates of the given scope for
+        # the given field.
+        self.assertQuerysetEqual(
+            Article.objects.dates('pub_date', 'year'),
+            ["datetime.datetime(2005, 1, 1, 0, 0)"])
+        self.assertQuerysetEqual(
+            Article.objects.dates('pub_date', 'month'),
+            ["datetime.datetime(2005, 7, 1, 0, 0)"])
+        self.assertQuerysetEqual(
+            Article.objects.dates('pub_date', 'day'),
+            ["datetime.datetime(2005, 7, 28, 0, 0)",
+             "datetime.datetime(2005, 7, 29, 0, 0)",
+             "datetime.datetime(2005, 7, 30, 0, 0)",
+             "datetime.datetime(2005, 7, 31, 0, 0)"])
+        self.assertQuerysetEqual(
+            Article.objects.dates('pub_date', 'day', order='ASC'),
+            ["datetime.datetime(2005, 7, 28, 0, 0)",
+             "datetime.datetime(2005, 7, 29, 0, 0)",
+             "datetime.datetime(2005, 7, 30, 0, 0)",
+             "datetime.datetime(2005, 7, 31, 0, 0)"])
+        self.assertQuerysetEqual(
+            Article.objects.dates('pub_date', 'day', order='DESC'),
+            ["datetime.datetime(2005, 7, 31, 0, 0)",
+             "datetime.datetime(2005, 7, 30, 0, 0)",
+             "datetime.datetime(2005, 7, 29, 0, 0)",
+             "datetime.datetime(2005, 7, 28, 0, 0)"])
+
+        # dates() requires valid arguments.
+        self.assertRaisesRegexp(
+            TypeError,
+            "dates\(\) takes at least 3 arguments \(1 given\)",
+            Article.objects.dates,
+        )
+
+        self.assertRaisesRegexp(
+            FieldDoesNotExist,
+            "Article has no field named 'invalid_field'",
+            Article.objects.dates,
+            "invalid_field",
+            "year",
+        )
+
+        self.assertRaisesRegexp(
+            AssertionError,
+            "'kind' must be one of 'year', 'month' or 'day'.",
+            Article.objects.dates,
+            "pub_date",
+            "bad_kind",
+        )
+
+        self.assertRaisesRegexp(
+            AssertionError,
+            "'order' must be either 'ASC' or 'DESC'.",
+            Article.objects.dates,
+            "pub_date",
+            "year",
+            order="bad order",
+        )
+
+        # Use iterator() with dates() to return a generator that lazily
+        # requests each result one at a time, to save memory.
+        dates = []
+        for article in Article.objects.dates('pub_date', 'day', order='DESC').iterator():
+            dates.append(article)
+        self.assertEqual(dates, [
+            datetime(2005, 7, 31, 0, 0),
+            datetime(2005, 7, 30, 0, 0),
+            datetime(2005, 7, 29, 0, 0),
+            datetime(2005, 7, 28, 0, 0)])
+
+        # You can combine queries with & and |.
+        s1 = Article.objects.filter(id__exact=a.id)
+        s2 = Article.objects.filter(id__exact=a2.id)
+        self.assertQuerysetEqual(s1 | s2,
+            ["<Article: Area man programs in Python>",
+             "<Article: Second article>"])
+        self.assertQuerysetEqual(s1 & s2, [])
+
+        # You can get the number of objects like this:
+        self.assertEqual(len(Article.objects.filter(id__exact=a.id)), 1)
+
+        # You can get items using index and slice notation.
+        self.assertEqual(Article.objects.all()[0], a)
+        self.assertQuerysetEqual(Article.objects.all()[1:3],
+            ["<Article: Second article>", "<Article: Third article>"])
+
+        s3 = Article.objects.filter(id__exact=a3.id)
+        self.assertQuerysetEqual((s1 | s2 | s3)[::2],
+            ["<Article: Area man programs in Python>",
+             "<Article: Third article>"])
+
+        # Slicing works with longs.
+        self.assertEqual(Article.objects.all()[0L], a)
+        self.assertQuerysetEqual(Article.objects.all()[1L:3L],
+            ["<Article: Second article>", "<Article: Third article>"])
+        self.assertQuerysetEqual((s1 | s2 | s3)[::2L],
+            ["<Article: Area man programs in Python>",
+             "<Article: Third article>"])
+
+        # And can be mixed with ints.
+        self.assertQuerysetEqual(Article.objects.all()[1:3L],
+            ["<Article: Second article>", "<Article: Third article>"])
+
+        # Slices (without step) are lazy:
+        self.assertQuerysetEqual(Article.objects.all()[0:5].filter(),
+            ["<Article: Area man programs in Python>",
+             "<Article: Second article>",
+             "<Article: Third article>",
+             "<Article: Article 6>",
+             "<Article: Default headline>"])
+
+        # Slicing again works:
+        self.assertQuerysetEqual(Article.objects.all()[0:5][0:2],
+            ["<Article: Area man programs in Python>",
+             "<Article: Second article>"])
+        self.assertQuerysetEqual(Article.objects.all()[0:5][:2],
+            ["<Article: Area man programs in Python>",
+             "<Article: Second article>"])
+        self.assertQuerysetEqual(Article.objects.all()[0:5][4:],
+            ["<Article: Default headline>"])
+        self.assertQuerysetEqual(Article.objects.all()[0:5][5:], [])
+
+        # Some more tests!
+        self.assertQuerysetEqual(Article.objects.all()[2:][0:2],
+            ["<Article: Third article>", "<Article: Article 6>"])
+        self.assertQuerysetEqual(Article.objects.all()[2:][:2],
+            ["<Article: Third article>", "<Article: Article 6>"])
+        self.assertQuerysetEqual(Article.objects.all()[2:][2:3],
+            ["<Article: Default headline>"])
+
+        # Using an offset without a limit is also possible.
+        self.assertQuerysetEqual(Article.objects.all()[5:],
+            ["<Article: Fourth article>",
+             "<Article: Article 7>",
+             "<Article: Updated article 8>"])
+
+        # Also, once you have sliced you can't filter, re-order or combine
+        self.assertRaisesRegexp(
+            AssertionError,
+            "Cannot filter a query once a slice has been taken.",
+            Article.objects.all()[0:5].filter,
+            id=a.id,
+        )
+
+        self.assertRaisesRegexp(
+            AssertionError,
+            "Cannot reorder a query once a slice has been taken.",
+            Article.objects.all()[0:5].order_by,
+            'id',
+        )
+
+        try:
+            Article.objects.all()[0:1] & Article.objects.all()[4:5]
+            self.fail('Should raise an AssertionError')
+        except AssertionError, e:
+            self.assertEqual(str(e), "Cannot combine queries once a slice has been taken.")
+        except Exception, e:
+            self.fail('Should raise an AssertionError, not %s' % e)
+
+        # Negative slices are not supported, due to database constraints.
+        # (hint: inverting your ordering might do what you need).
+        try:
+            Article.objects.all()[-1]
+            self.fail('Should raise an AssertionError')
+        except AssertionError, e:
+            self.assertEqual(str(e), "Negative indexing is not supported.")
+        except Exception, e:
+            self.fail('Should raise an AssertionError, not %s' % e)
+
+        error = None
+        try:
+            Article.objects.all()[0:-5]
+        except Exception, e:
+            error = e
+        self.assertTrue(isinstance(error, AssertionError))
+        self.assertEqual(str(error), "Negative indexing is not supported.")
+
+        # An Article instance doesn't have access to the "objects" attribute.
+        # That's only available on the class.
+        self.assertRaisesRegexp(
+            AttributeError,
+            "Manager isn't accessible via Article instances",
+            getattr,
+            a7,
+            "objects",
+        )
+
+        # Bulk delete test: How many objects before and after the delete?
+        self.assertQuerysetEqual(Article.objects.all(),
+            ["<Article: Area man programs in Python>",
+             "<Article: Second article>",
+             "<Article: Third article>",
+             "<Article: Article 6>",
+             "<Article: Default headline>",
+             "<Article: Fourth article>",
+             "<Article: Article 7>",
+             "<Article: Updated article 8>"])
+        Article.objects.filter(id__lte=a4.id).delete()
+        self.assertQuerysetEqual(Article.objects.all(),
+            ["<Article: Article 6>",
+             "<Article: Default headline>",
+             "<Article: Article 7>",
+             "<Article: Updated article 8>"])
+
+    @skipUnlessDBFeature('supports_microsecond_precision')
+    def test_microsecond_precision(self):
+        # In PostgreSQL, microsecond-level precision is available.
+        a9 = Article(
+            headline='Article 9',
+            pub_date=datetime(2005, 7, 31, 12, 30, 45, 180),
+        )
+        a9.save()
+        self.assertEqual(Article.objects.get(pk=a9.pk).pub_date,
+            datetime(2005, 7, 31, 12, 30, 45, 180))
+
+    @skipIfDBFeature('supports_microsecond_precision')
+    def test_microsecond_precision_not_supported(self):
+        # In MySQL, microsecond-level precision isn't available. You'll lose
+        # microsecond-level precision once the data is saved.
+        a9 = Article(
+            headline='Article 9',
+            pub_date=datetime(2005, 7, 31, 12, 30, 45, 180),
+        )
+        a9.save()
+        self.assertEqual(Article.objects.get(id__exact=a9.id).pub_date,
+            datetime(2005, 7, 31, 12, 30, 45))
+
+    def test_manually_specify_primary_key(self):
+        # You can manually specify the primary key when creating a new object.
+        a101 = Article(
+            id=101,
+            headline='Article 101',
+            pub_date=datetime(2005, 7, 31, 12, 30, 45),
+        )
+        a101.save()
+        a101 = Article.objects.get(pk=101)
+        self.assertEqual(a101.headline, u'Article 101')
+
+    def test_create_method(self):
+        # You can create saved objects in a single step
+        a10 = Article.objects.create(
+            headline="Article 10",
+            pub_date=datetime(2005, 7, 31, 12, 30, 45),
+        )
+        self.assertEqual(Article.objects.get(headline="Article 10"), a10)
+
+    def test_year_lookup_edge_case(self):
+        # Edge-case test: A year lookup should retrieve all objects in
+        # the given year, including Jan. 1 and Dec. 31.
+        a11 = Article.objects.create(
+            headline='Article 11',
+            pub_date=datetime(2008, 1, 1),
+        )
+        a12 = Article.objects.create(
+            headline='Article 12',
+            pub_date=datetime(2008, 12, 31, 23, 59, 59, 999999),
+        )
+        self.assertQuerysetEqual(Article.objects.filter(pub_date__year=2008),
+            ["<Article: Article 11>", "<Article: Article 12>"])
+
+    def test_unicode_data(self):
+        # Unicode data works, too.
+        a = Article(
+            headline=u'\u6797\u539f \u3081\u3050\u307f',
+            pub_date=datetime(2005, 7, 28),
+        )
+        a.save()
+        self.assertEqual(Article.objects.get(pk=a.id).headline,
+            u'\u6797\u539f \u3081\u3050\u307f')
+
+    def test_hash_function(self):
+        # Model instances have a hash function, so they can be used in sets
+        # or as dictionary keys. Two models compare as equal if their primary
+        # keys are equal.
+        a10 = Article.objects.create(
+            headline="Article 10",
+            pub_date=datetime(2005, 7, 31, 12, 30, 45),
+        )
+        a11 = Article.objects.create(
+            headline='Article 11',
+            pub_date=datetime(2008, 1, 1),
+        )
+        a12 = Article.objects.create(
+            headline='Article 12',
+            pub_date=datetime(2008, 12, 31, 23, 59, 59, 999999),
+        )
+
+        s = set([a10, a11, a12])
+        self.assertTrue(Article.objects.get(headline='Article 11') in s)
+
+    def test_extra_method_select_argument_with_dashes_and_values(self):
+        # The 'select' argument to extra() supports names with dashes in
+        # them, as long as you use values().
+        a10 = Article.objects.create(
+            headline="Article 10",
+            pub_date=datetime(2005, 7, 31, 12, 30, 45),
+        )
+        a11 = Article.objects.create(
+            headline='Article 11',
+            pub_date=datetime(2008, 1, 1),
+        )
+        a12 = Article.objects.create(
+            headline='Article 12',
+            pub_date=datetime(2008, 12, 31, 23, 59, 59, 999999),
+        )
+
+        dicts = Article.objects.filter(
+            pub_date__year=2008).extra(
+                select={'dashed-value': '1'}
+            ).values('headline', 'dashed-value')
+        self.assertEqual([sorted(d.items()) for d in dicts],
+            [[('dashed-value', 1), ('headline', u'Article 11')], [('dashed-value', 1), ('headline', u'Article 12')]])
+
+    def test_extra_method_select_argument_with_dashes(self):
+        # If you use 'select' with extra() and names containing dashes on a
+        # query that's *not* a values() query, those extra 'select' values
+        # will silently be ignored.
+        a10 = Article.objects.create(
+            headline="Article 10",
+            pub_date=datetime(2005, 7, 31, 12, 30, 45),
+        )
+        a11 = Article.objects.create(
+            headline='Article 11',
+            pub_date=datetime(2008, 1, 1),
+        )
+        a12 = Article.objects.create(
+            headline='Article 12',
+            pub_date=datetime(2008, 12, 31, 23, 59, 59, 999999),
+        )
+
+        articles = Article.objects.filter(
+            pub_date__year=2008).extra(
+                select={'dashed-value': '1', 'undashedvalue': '2'})
+        self.assertEqual(articles[0].undashedvalue, 2)