r13085 introduced a couple of tests that fail when run under Python 2.3:
C:\u\kmt\django\branch1.1.X\tests>"\bin\Python 23\python.exe" runtests.py --settings=testdb.sqlite utils
======================================================================
FAIL: Modules deep inside an egg can still be tested for existence
----------------------------------------------------------------------
Traceback (most recent call last):
File "C:\u\kmt\django\branch1.1.X\tests\regressiontests\utils\module_loading.py", line 67, in test_deep_loader
self.assertRaises(ImportError, import_module, 'egg_module.sub1.sub2.bad_module')
File "C:\bin\Python23\lib\unittest.py", line 295, in failUnlessRaises
raise self.failureException, excName
AssertionError: ImportError
======================================================================
FAIL: Module existence can be tested inside eggs
----------------------------------------------------------------------
Traceback (most recent call last):
File "C:\u\kmt\django\branch1.1.X\tests\regressiontests\utils\module_loading.py", line 48, in test_shallow_loader
self.assertRaises(ImportError, import_module, 'egg_module.bad_module')
File "C:\bin\Python23\lib\unittest.py", line 295, in failUnlessRaises
raise self.failureException, excName
AssertionError: ImportError
----------------------------------------------------------------------
Ran 36 tests in 0.109s
FAILED (failures=2)
We've previously "fixed" (by simply skipping the tests under Python 2.3) a couple of problems where Python 2.3 won't raise an ImportError on an attempt to re-import a module that raises an error (see r13022, r13032), but I have not had time to figure out if that is what is happening here or if there is something else going on.
Possible fix for Python2.3 failures