This abuses errors in the fixed-point math, to have the first “error” line up exactly where we want, on that final leap second.A round-down multiplier is used, with a pre-increment, as that is the type of adjustment that yields an eventual long-minute and long-hour.

i love it