fix(tests): it only runs when both `blocking` and `async` features are active

[oleonardolima]

Describe the enhancement

As of working and debugging some PRs, I noticed that basically all of the tests only run when both blocking and async features are active, which is great in an adversarial scenario to guarantee that both are producing the same expected results given the same API usage.

However, the CI steps for using specific features, either blocking or async seem to only build the code, as the test is filtered out.

I think the tests should still run when either of blocking or async features are active.

Use case

Execute the test scenarios on every variation of the CI matrix.

Additional context