ModInt overflow issue

[plasmatic1]

templates/math/mod.hpp

Line 20 in de5c9c1

template <typename U> static T normalize(const U& x) {

assumes that the initial value of U can be held in T. This is not always the case (i.e. when U is int64_t and T is int32_t). To fix this, change the line to T res = x % mod();