v = ','.join(v) # Quirk: must format list value manually.

        Setting.create_or_update(k, v, t)

    session.commit()

@pytest.fixture

def auto_clear_ip_permissions():

    """Fixture that provides nothing but clearing IP permissions upon test

    exit. This clearing is needed to avoid other test failing to make fake http

    accesses."""

    yield

    # cleanup

    user_model = UserModel()

    user_ids = []

    user_ids.append(User.get_default_user().user_id)

    user_ids.append(User.get_by_username(TEST_USER_REGULAR_LOGIN).user_id)

    for user_id in user_ids:

        for ip in UserIpMap.query().filter(UserIpMap.user_id == user_id):

            user_model.delete_extra_ip(user_id, ip.ip_id)

    # IP permissions are cached, need to invalidate this cache explicitly

    invalidate_all_caches()

@pytest.fixture

def test_context_fixture(app_fixture):

    """

    Encompass the entire test using this fixture in a test_context,

    making sure that certain functionality still works even if no call to

    self.app.get/post has been made.

    The typical error message indicating you need a test_context is:

        TypeError: No object (name: context) has been registered for this thread

    The standard way to fix this is simply using the test_context context

    manager directly inside your test:

        with test_context(self.app):

            <actions>

    but if test setup code (xUnit-style or pytest fixtures) also needs to be

    executed inside the test context, that method is not possible.

    Even if there is no such setup code, the fixture may reduce code complexity

    if the entire test needs to run inside a test context.

    To apply this fixture (like any other fixture) to all test methods of a

    class, use the following class decorator:

        @pytest.mark.usefixtures("test_context_fixture")

        class TestFoo(TestController):