[DPE-4068] Convert all tests from unittest to pytest + enable/disable secrets in every test #452
Conversation
| Typically: to save CI time, when the same check were executed in a Juju 3-specific way already | ||
| """ | ||
| if juju_has_secrets: | ||
| pytest.skip("Skipping legacy secrets tests") |
There was a problem hiding this comment.
We had already the whole test suite running with and without secrets (unit-juju2 and unit-juju3 respectively), the only change needed was to adapt the few tests that only ran in one or the other, and remove this fixtures.
There was a problem hiding this comment.
merging those 2 suites into one will be addressed in a follow-up PR!
| @patch_network_get(private_address="1.1.1.1") | ||
| @patch("charm.PostgresqlOperatorCharm._on_leader_elected") | ||
| @pytest.mark.usefixtures("use_caplog") | ||
| def test_delete_existing_password_secrets(self, _, __): |
There was a problem hiding this comment.
similar to what was done in canonical/postgresql-operator#425, this test got removed and test_delete_password runs with and without secrets
| ): | ||
| # as this test checks for a migration from databag to secrets, | ||
| # there's no need for this test when secrets are not enabled. | ||
| if not juju_has_secrets: |
There was a problem hiding this comment.
here and on the test below, same idea from canonical/postgresql-operator#425
| tc.assertEqual(len(first_password), 16) | ||
| tc.assertIsNotNone(re.fullmatch("[a-zA-Z0-9\b]{16}$", first_password)) |
There was a problem hiding this comment.
We can just use regular asserts in such cases. No need to keep the test case just for that.
There was a problem hiding this comment.
agreed, will refactor this together with merging the 2 suites in the follow-up PR.
This reverts commit 9137e67.
… secrets in every test (canonical#452) * convert all tests to pytest * remove secret remove hook * Revert "remove secret remove hook"
Follow-up to DPE-2674 in canonical/postgresql-operator#425