Fixes #169
MSVC issues a warning when we try to negate an unsigned number. The goal
however was to perform bit twiddling, so I used bitwise negation.
Hopefully the compiler will not complain about overflow, which is
perfectly well defined on unsigned numbers.
carry += (u64)t[i+8] + ~l[i];
carry >>= 32;
}
- u32 mask = (u32)-carry; // carry == 0 or 1
+ u32 mask = ~(u32)carry + 1; // carry == 0 or 1
FOR (i, 0, 8) {
carry += (u64)t[i+8] + (~l[i] & mask);
u[i] = (u32)carry;