I can't imagine a way that this can work but thought I'd post a feature request to see if bigger brains than mine can come up with anything.
I wrote a great PoC with cats-effect-testing with a test like:
"foo" in IO {
"stuff"
}: IO[String]
The problem is in reality most of our tests use Fixture Contexts to mix in fixture data, so would look like:
"foo" in new TestFixture with OtherTestFixture {
IO {
"stuff"
}
}: TestFixture with OtherTestFixture
The problem of course is : TestFixture with OtherTestFixture which means the test doesn't return an IO[_] which means the library can't evaluate it.
I can't think of a way to both keep the use of fixture contexts to minimise change, but also be able to have tests that must be effectful (i.e. Acceptance/E2E tests)
I can't imagine a way that this can work but thought I'd post a feature request to see if bigger brains than mine can come up with anything.
I wrote a great PoC with cats-effect-testing with a test like:
The problem is in reality most of our tests use Fixture Contexts to mix in fixture data, so would look like:
The problem of course is
: TestFixture with OtherTestFixturewhich means the test doesn't return anIO[_]which means the library can't evaluate it.I can't think of a way to both keep the use of fixture contexts to minimise change, but also be able to have tests that must be effectful (i.e. Acceptance/E2E tests)