Bug: some failure tests are wrongly set up and not testing intended behavior

[crowecawcaw]

Describe the bug

Some unit tests that verify validation failures have mistakes in the test setup (e.g. using the wrong model) but the test only looks for a validation error so the test still passes. As a result, these tests do not actually verify the library behavior.

See this test which mixes attributes and amounts:

openjd-model-for-python/test/openjd/model/v2023_09/test_step_host_requirements.py

Lines 268 to 282 in 96651fa

	def test_non_standard_attribute_capability_noncompliant_value_string(
	self, field: str, value: str, error_count: int
	) -> None:
	# Test the constraints on an amount capability value within the
	# anyOf and allOf clauses raise validation errors when they're violated.
	# GIVEN
	data = {"name": "attr.custom", field: [value]}
	# WHEN
	with pytest.raises(ValidationError) as excinfo:
	_parse_model(model=AmountRequirementTemplate, obj=data)
	# THEN
	assert len(excinfo.value.errors()) == error_count, str(excinfo.value)

See this PR which fixes a test that mixes HostRequirements and AmountRequirements models: https://github.com/OpenJobDescription/openjd-model-for-python/pull/265/changes

Recommend setting a new pattern for testing for validation failures that causes a test to fail if the wrong model is used, then applying that patter across the codebase. See the PR as a possible solution.

Expected Behaviour

...

Current Behaviour

...

Reproduction Steps

...

Environment

...