Changeset 7789 for django/trunk/tests/regressiontests/fixtures_regress

Timestamp:

06/29/08 19:38:14 (5 months ago)

Author:

mtredinnick

Message:

Fixed #7565 -- Fixed a problem with PostgreSQL sequence resetting in loaddata.

Files:

• django/trunk/tests/regressiontests/fixtures_regress/fixtures/model-inheritance.json (added)
• django/trunk/tests/regressiontests/fixtures_regress/models.py (modified) (5 diffs)

django/trunk/tests/regressiontests/fixtures_regress/models.py

@@ -21 +21 @@
-    name = models.CharField(max_length=20, null=True)
-    owner = models.ForeignKey(User, null=True)
-
+
-    def __unicode__(self):
-        # Oracle doesn't distinguish between None and the empty string.
@@ -39 +39 @@
-        Absolute.load_count += 1
-
+class Parent(models.Model):
+    name = models.CharField(max_length=10)
+
+class Child(Parent):
+    data = models.CharField(max_length=10)
+
-
-__test__ = {'API_TESTS':"""
@@ -45 +51 @@
-# Load a fixture that uses PK=1
->>> management.call_command('loaddata', 'sequence', verbosity=0)
-
+
-# Create a new animal. Without a sequence reset, this new object
-# will take a PK of 1 (on Postgres), and the save will fail.
@@ -62 +68 @@
-
-###############################################
- 
+
-# os.path.join will throw away the initial parts of a path if it encounters
- 
+
-# we need to make sure we don't discover the absolute path in every fixture directory.
-
@@ -95 +101 @@
->>> sys.stderr = savestderr
-
+###############################################
+# Test for ticket #7565 -- PostgreSQL sequence resetting checks shouldn't
+# ascend to parent models when inheritance is used (since they are treated
+# individually).
+
+>>> management.call_command('loaddata', 'model-inheritance.json', verbosity=0)
+
-"""}