mirror of
https://github.com/boostorg/random.git
synced 2026-01-19 04:22:17 +00:00
Multiprecision now performs correct rounding when converting to double - previously the max value for uint1024_t would just fit in a double if it was truncated. But now that it's correctly rounded it overflows and breaks the calculation.
7.9 KiB
7.9 KiB