fix(model): make Secret.set_content invalidate local cache #1001
Conversation
This is so that if get_content() is called after, it fetches and returns the expected new content. We could also set self._content to the new content, but re-fetching it from Juju seems better.
| old_content = secret.get_content() | ||
| self.assertEqual(old_content, {'foo': 'bar'}) | ||
| secret.set_content({'new': 'content'}) | ||
| fake_script(self, 'secret-get', """echo '{"new": "content"}'""") |
There was a problem hiding this comment.
I wish we didn't have to force the fake_script, that certainly makes the test feel like a tautology (though I realize it actually isn't).
Also, should we be adding a test around testing.Harness and that when we're using its backend that we get the right behavior?
There was a problem hiding this comment.
Yeah, the second fake_script felt weird, but it's correct here. I thought about have a more realistic secret-set/secret-get that wrote to and read from a file instead, but that seems like unnecessary work for what we're testing here. Here we just need to assert that the second get_content() call returns the updated content rather than the cached content as it did previously.
Regarding Harness, I think that is an interesting idea in general, but I don't think we'd be adding anything for this case by adding a specific test. This isn't a Harness problem but a model.Secret problem -- it's backend-independent.
This is so that if get_content() is called after, it fetches and returns the expected new content. We could also set self._content to the new content, but re-fetching it from Juju seems better.