Dominik Kellner

A better workaround might be to use `patch`: ```python from unittest.mock import patch @patch('sys.argv', ['./manage.py', 'dumpdata']) def dump_json(self): call_command('dumpdata', '--indent=2', '--output=dump.json') ```

@funkyfuture The reasoning behind `test_default_none_nullable` and `test_default_setter_none_nullable` is that if a field is nullable and a document already has `None` set for this field, then normalization (with a `default` value...

@funkyfuture You're delegating to Vietnam here, so much for the heat ;-). I will take a look at it and get back to you until wednesday.

@funkyfuture Just to be sure, you want the behaviour you described before, right? So far I've refactored the tests using `mark.parametrize` and fixed `test_default_none_not_nullable` as we talked before. > -...

Yes, it was in PR #282. The other (non-simple) fix we had for that is in #272, and I still have the branch here: https://github.com/dkellner/cerberus/tree/default-readonly . IIRC and Cerberus still...

> anyway, has someone the time to tackle this bug? it's the hardest nut on the 1.3 playlist so far. I'm afraid I will not find the time for that...

Thanks for the detailed report! Do you have time to give this a shot? Your example looks like a good starting point for an integration test in the spirit of...

I think for now the best way to address this would be using events: http://python-eve.org/features.html#event-hooks. Since Cerberus 1.0 supports a `default_setter` rule, maybe this will be another way to tackle...

@hectorcanto Any help is highly appreciated! At first you could write a new integration test for using the SQLAlchemy default function - maybe even for different types of fields. Since...

@ktal90 Just to be clear, you wish to support something like the following, to create a new parent along with two new children (based on `examples/one_to_many`), right? ``` $ curl...