Hojun-Cho
69e31355ce
i64dec computed the magnitude with `n = -n` inside i64, which wraps at i64::MIN (-MIN == MIN stays negative), so the `for (n > 0)` loop never ran and println(i64::MIN) emitted only the bare '-'. The whole print/ println/fprint family was affected at the single boundary value (and int MIN, since int is 8B). ref/hare/fmt/print.ha:124-129 takes the magnitude through math::absi64 into a u64; do the same (math.absi64 exists, lib/math/math.ww:12). ww's own printf path was already correct ((-v): u64), so print and printf now agree. fmttest gains fprinti64min + fprintintmin pinning the boundary.