`<format>`: `format_to_n` missing difference type casting

This bug belongs to difference type casting category.

https://github.com/microsoft/STL/blob/4483e87f6e5aa5c4598404fc01eecb7e547dd439/stl/inc/format#L3551-L3557

The type of _Max is iter_difference_t<O>, but the parameter type of _Fmt_iterator_buffer is ptrdiff_t, which does not guarantee that it can be implicitly converted by the former.

contrived testcase:

#include <ranges>
#include <format>
#include <vector>
#include <iostream>

int main() {
  std::vector<char> v(10);
  auto r = std::views::iota(0ULL) | 
    std::views::transform([&](std::size_t i) -> auto& { return v[i]; });
  auto str = std::format_to_n(r.begin(), 5, "{}", "hello");
  std::cout << v.data() << "\n";
}

https://godbolt.org/z/5doen3P5d

Start at stl/inc/format around lines 3551-3557 and compare the _Max and _Fmt_iterator_buffer parameter types with the provided format_to_n testcase. Verify the corrected difference-type handling using that testcase and confirm the formatting behavior remains correct.