nvm, I missed the literal value injections. I still hold that it's a bad test since JS doesn't care about types of numbers so saying it must be a float is confusing, but I must have messed up something else.

Not sure, but letting dangling issues hanging around is not a good idea either.

The relevant test is this:

describe('Mean is returned as float', function(){

  tracker = new TempTracker();

  tracker.insert(1);
  tracker.insert(2);
  tracker.insert(4);

  var mean = tracker.getMean();
  Test.expect(Number(mean) && mean % 1 !== 0, 'getMean() didn\'t return a float. Returned: ' + mean);
});

So unless something is really messed up it shouldn't fail. It doesn't even check the result is the same as the expected value.

Well considering that JavaScript does not actually have floats, or more specifically everything is a float, (https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures), there is an issue with the construction of the tests or at least the phrasing of them. Since testing if something is a float is literally testing if it's a number, this test must be doing something else. Sadly, I have not used this site in nearly 3 years so don't know what I actually coded back then, but I assue you there was good reason for my frustration.

As a related question - I seem to recall this site being very new at that time. Was "question" even an option then vs. issue?

Your "issue" is useless and isn't really a issue (it's a question) either ;-)

Next time if you can't get through a specific part, mark it as a question and not an issue. I'm resolving it because it's not an issue, and I said no because the situation you're describing is not possible unless there are dumb mistakes in your code (I solved it so I can see the tests), and without your code that's the only possible response.

Remeber: 99% of the "issues" comes from your own end unless you're experienced ;-)

Impressively useless response! So glad that you were able to mark the issue "resolved" by doing that. /facepalm

No.

Why "79999997"? Nothing should happen there, just as with "78987".

Actually, running "797" repleace exactly twice will solve the test case 797979797979797979797979.

But I agree with you to add test case 799999997

If you try again.. are the tests changed or fixed? Also When I ran it it seems to work fine?

It errored out with "some is not a fuction". So unless there was a requirement I didn't see to implement something that had a function on it called "some" that the test could call... then the tests were broken.

Also, this was only the "user test" setion, the on-submit tests work fine (and, I notice, don't have the crazy setup).

(Edit to note that this is for the Javascript version. It's possible the tests were ported badly from a functional language or something.)

That's not an issue the function is used to evaluate the correct result from your one.

Something was really messed up with the provided "user tests". I couldn't make heads or tales of them, so just deleted them and used the submit ones.

Nice kata.

1 significant problem, as illustrated by the top solution - your tests only have false cases that are negative or even. You need far more tests to prove a valid function. Consider a random test generator paired with a known-to-be-accurate solution, combined with several non-even false primes.

Very unhelpful response. What is missing that you needed?