that's not the test that failed, your prints are misleading. if the rust tests are to be blamed in some way it's that they aren't separating the tests, which makes it easier to mix them up, and more generally that they invite mistrust by not showing what they're doing, relying instead on the solver to figure out what happened which isn't ideal when the test code is hidden

(also you absolutely should not be using f64, that would fail for large numbers since you're effectively reducing the number of available bits)