> Ok ... if the max value is outside the bounds and the code should have thrown
> an error, then I apologize.
No issue. Looking again at it I’m still not sure which bit or code or test is
wrong or perhaps neither(?) but I think it should be consistent. Perhaps
Sebastian has a better idea?
> But if you find something like this, wouldn't it be better to fix the broken
> code or at least leave a comment in the test that is guaranteed to break?
I wasn't 100% sure which method had the incorrect boundary condition.