Thread: How does one catch NaN type errors?

So, to catch NaN and infinity, maybe:

Code:

bool iFinite(double d)
{
    return (d != numeric_limits<double>::quiet_NaN() && d != numeric_limits<double>::infinity());
}

Or to be more thorough:

Code:

#include <limits>

template <typename T>
inline bool isFinite(T val)
{
    if (std::numeric_limits<T>::has_quiet_NaN && val == std::numeric_limits<T>::quiet_NaN())
        return false;
    if (std::numeric_limits<T>::has_infinity && val == std::numeric_limits<T>::infinity())
        return false;
    return true;
}

edit: Apparently this doesn't find indeterminate numbers (0.0/0.0).