pub fn u64_to_f64(v: u64) -> f64
u64 → f64 — always produces a valid f64, but may lose precision for v > 2^53.
u64 → f64
v > 2^53