I can't imagine flipping a hash with duplicate values ever really had a well-defined result in older perl versions, but the result is much more random in 5.18. I simply deleted the repeated 'bar' value in the test hash so all the values are unique, resulting in unique keys in the flipped hash.
I can't imagine flipping a hash with duplicate values ever really had a well-defined result in older perl versions, but the result is much more random in 5.18. I simply deleted the repeated 'bar' value in the test hash so all the values are unique, resulting in unique keys in the flipped hash.