Compare floats while safely handling invisible float precision overflow

epsilonCompare.hpp

#pragma once

#include <limits>

/**
* Check if a and b are within epsilon range of eachother.
*
* Based on this post: https://www.reddit.com/r/ProgrammerHumor/comments/ggwxjp/the_dwindling_sanity_of_valves_programmers/fq646zf?utm_source=share&utm_medium=web2x
**/
template<typename T>  // Numeric type
bool epsilonCompare(T a, T b)
{
    using epsilon = std::numeric_limits<double>::epsilon();
    return (a - epsilon < b && res + epsilon > b);
}