thanks.

i got a question. it's not regrading to the solution but about how to write test case. how do you proof it's a "true" random for each color generation in the test case? run test 1000 times and calculate each color count to 25% of total?

is that any easy way to do that in the test framework?

I think so too. I tried another kata with ruby. Running test already returned the same error, whereas directly submit final solution without any problem.

My ruby solution passed irb locally (irb 0.9.6 & 1.9.3p484), but I keep getting this error:

Test didn't pass: Unknown error

Since I have done python version before & passed, I believe it should not be my coding issue. Is there anybody could take a look?

Thanks.