fork approved, issue fixed

Java Fork

• makes KM_PER_MILE and LITRES_PER_GALLON doubles as suggested by siebenschlaefer 7 years ago
• makes reference solution static
• ups number of random tests to 100
• silences a deprecated API warning by removing java.math.BigDecimal
• (had to replace reference solution)
• adds edge case found by kayakero 8 years ago to sample and fixed tests

Or is there a way to do it in java without using math.round or DecimalFormat?

@siebenschlaefer does this mean that this one can't be completed in java until it is fixed? My quest to finish all the 8kyu in java is going down the drain haha

Not unless the kata explicitly asks you to solve the problem without using this very method.

I could reproduce this issue.
The Java tests use floats for the constants. So the expected value before rounding of mpgToKPM(4145) is calculated as 1467.35498046875 (instead of 1467.35505046589605626650154590606689453125 if doubles were used).

@emporio, @AngusJGoldsmith or anyone with write access, please fix this by replacing

  private static float KM_PER_MILE = 1.609344f;
  private static float LITRES_PER_GALLON = 4.54609188f;

in the tests with

  private static double KM_PER_MILE = 1.609344;
  private static double LITRES_PER_GALLON = 4.54609188;

But that would really defeat the concept of doing the kata, wouldn't it?

@emporio, I think the solution (in Java) is something along the lines of (where n is a double or float) (n*100)/100.00

Can someone proficient in Java repair this?

Reusing variables is not as clear as possible. And assigning to a reference variable doesn't create a copy.

No, it isn't. If so, you can declare your parameter final.

I just didn't want to add a new String() to make a copy. I wanted it to stay as clear as possible.