GeneROMst: Incorect generational index normalization formula for wrap-around

[eddyp]

Current description says "in case of wrap around, to ease up chronological ordering, we need to calculate normalized generation indexes of entry i (gi): gi = (Gi + N) % N",

But looking at the example values for wrap-around case 2 3 5 18 , and the example gi values for these (18 19 21 2), the applied formula was actually gi = (Gi + N) % 2N.

Note that if the '% N' formula is applied, we'd get collisions on the example between Gi=2 and Gi=18, as they'd map to gi=2.

Hence the right formula should be %2N.

The correctness of the normalization could be checked based on the property "starting with some arbitrary Gi in [0, 2N), write an arbitrary number of entries <= N, and gi (the normalized Gi) should be oldest/smallest"