(lhs: Rational, rhs: Rational)
| 60 | * so every ordinary value keeps the same double bit for bit. (A bigint part |
| 61 | * between `2^53` and the largest double is rounded when it is converted, so |
| 62 | * that quotient is rounded twice and can be one unit in the last place from |
| 63 | * the nearest double; this is kept so the doubles do not change.) |
| 64 | * |
| 65 | * A bigint numerator or denominator can be larger than the largest double |
| 66 | * (about `1.8·10^308`). Its `Number()` is then `Infinity`, and the quotient |
| 67 | * of the two conversions is wrong: `(10^400 + 1)/10^400` gave |
no test coverage detected