On Thu, Sep 5, 2024 at 4:07 AM Juraj Linkeš <juraj.lin...@pantheon.tech> wrote: > > > > On 26. 8. 2024 18:50, Jeremy Spewock wrote: > > On Wed, Aug 21, 2024 at 10:53 AM Juraj Linkeš > > <juraj.lin...@pantheon.tech> wrote: > > <snip> > >> class DTSRunner: > >> @@ -232,9 +231,9 @@ def _get_test_suites_with_cases( > >> > >> for test_suite_config in test_suite_configs: > >> test_suite_class = > >> self._get_test_suite_class(test_suite_config.test_suite) > >> - test_cases = [] > >> - func_test_cases, perf_test_cases = self._filter_test_cases( > >> - test_suite_class, test_suite_config.test_cases > >> + test_cases: list[type[TestCase]] = [] > > > > If TestCase is just a class, why is the `type[]` in the annotation > > required? Are these not specific instances of the TestCase class? I > > figured they would need to be in order for you to run the specific > > test case methods. Maybe this has something to do with the class being > > a Protocol? > > > > The *_test decorators return type[TestCase]. The functions (test > methods) are cast to type[TestCase] (which kinda makes them subclasses > of TestCase).
Oh interesting, I didn't make the connection that casting them to type[TestCase] was similar to having them be subclasses of the type, but this actually makes a lot of sense. Thank you for the clarification! > > This was a suggestion from Luca and I took it as as. Maybe the functions > could be cast as instances of TestCase, but I didn't try that. Right, I would think that they could be cast directly to it, but there's no need obviously so that makes sense. > > >> + func_test_cases, perf_test_cases = > >> test_suite_class.get_test_cases( > >> + test_suite_config.test_cases > >> ) > >> if func: > >> test_cases.extend(func_test_cases) <snip> >
