Some new-class tests were added, and Rizvanov++ de xXx's solutions failed to pass them. Anyway these solutions are quite effective, more over, we were surprised to see such approach. We believe Rizvanov++ de xXx will be able to find some more optimizations and pass the TL.

We are sorry for the rejudges, but we will continue to support this problem, to add new tests and to rejudge it - until true solutions only could pass the TL and get AC.