repeated "No fixtures found." on manage.py test

[Waldemar Kornewald]

We're using Django's TestCase class when we run manage.py test this happens (desired behavior: the "No fixtures found." message should only be printed once - if it's needed at all):

No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
.No fixtures found.
...
----------------------------------------------------------------------
Ran 44 tests in 4.916s

OK

[Gabriel Hurley]

I've had no luck duplicating this so far... I've tried having no fixtures, empty fixtures, etc. and I've tried it with a single TestCase class, multiple TestCase classes, multiple test methods on those classes, different verbosity levels... Every way I try it I only get the "no fixtures found." method once.

I may be missing something. But for the sake of asking, are you doing something in the setUp or tearDown method that might be triggering this?

Other thought: could this be a DB-specific problem where a fixture is being re-loaded each time instead of a rollback? My attempts at duplicating this bug are simply on sqlite.

[Russell Keith-Magee]

I haven't ever seen this. You'll need to provide a test case to demonstrate how and when this happens.

[Jacob]

Milestone 1.2 deleted