Prevent integer overflow in FindAutomorphismIndex2nComplex()

[dsuponitskiy]

1. Changed the type of g0 and g from int32_t to uint64_t:
   • to prevent integer overflow if the result of (g*g0) is too high
   • NativeInteger().ModInverse() always returns positive number, so the type can be unsigned
2. Replaced "if (i == int32_t(m - 1))" with "if (i == (static_cast<int32_t>(m) - 1))" as the subtraction should be performed on int32_t to avoid unexpected behavior.
3. Replaced the C-style cast with the C++ one