在linux和windows上的最小/最大类型演绎不同。

In the following code, the template type deduction for std::min/max seems to be weird and I'd like to know why and how to properly fix it. In the following code, the template type deduct