Question

Well i’m a ruby newbie and im trying to learn with RubyKoans but i got stucked with this test

def test_dice_values_should_change_between_rolls
 48     dice = DiceSet.new
 49     dice.roll(5)
 50     first_time = dice.values
 51    
 52     dice.roll(5)
 53     second_time = dice.values
 54     
 55     assert_not_equal first_time, second_time,
 56       "Two rolls should not be equal"
 57   end

and this is DiceSet class

5  class DiceSet
  6    attr_accessor :values
  7 ··
  8    def initialize
  9      @values = []
 10    end
 11 
 12    def roll(times)
 13      @values.clear
 14      times.times do |x|
 15        @values << ( 1 + rand(6))
 16      end
 17     end
 18 ····
 19    end

the thing here is that whenever i run the code it always generates the exact same set of numbers, this is the Output.

Two rolls should not be equal.  <[3, 2, 4, 1, 3]> expected to be != to  <[3, 2, 4, 1, 3]>.

in the test im calling DiceSet.roll two times and for those two times i get the exact same set of ‘random’ numbers when they’re supossed to be diferent right? I figured that i just might create another instance of DiceSet in order to pass the test but im guessing that is not the objective of the test

Answer 1

The problem is that DiceSet#values returns a reference to an array, and that array stays the same for the whole lifetime of your DiceSet object. In DiceSet#roll you clear that array and then add new numbers. Since both calls to DiceSet#values return the same reference, the result of the first roll will be lost, and your test is comparing the array with itself.

I am not familiar with the RubyKoans and what requirements they have, i.e. if you DiceSet is supposed to store the values etc. If it is, then the most straightforward solution is to either use two DiceSets or use Object#dup to store a copy of the returned object for the test.

Be aware, however, that your test is fragile even with correctly functioning code, as there always is the chance that two consecutive rolls will return the exact same numbers. In this particular case it is relatively small but still very much existent.