Thanks! Good that old solutions still work.

Should be fixed now. The tests are essentially the same, but the fractions get now fully reduced before they're handed to the solver. That way, old solutions still work, the reference solution stays the same.

Sure: you generate a ratio and reduce it before you apply the tests.

I'll have a look on Sunday evening/Monday, but here's what's wrong: the current reference solution only takes the lowest common denominator of all denominators and doesn't check whether all ratios can be reduced to an even smaller denominator. A fix should be rather easy, but it's cumbersome without a proper keyboard.

@bkaes: Can you have a look?

In Haskell, I am getting a lot of tests failing on what seems to be (to me ;)) correct results.

For example:

expected: [(24,12),(36,12),(24,12),(12,12),(16,12),(12,12),(18,12)]
but got: [(12,6),(18,6),(12,6),(6,6),(8,6),(6,6),(9,6)]

Input is:
[(4,2),(6,2),(4,2),(2,2),(4,3),(2,2),(6,4)]

Am I misssing something? My solution is equivalent to the expected case, but the denominator is smaller.

Thanks!