Possible logic error in Floating-point Comparison function

[linomp]

In 02.Basic_Concepts_I.pdf, on slide 94 inside the function areFloatNearlyEqual there is the following code for catching Inf or NaN arguments:

if (std::isfinite(a) || isfinite(b))  // a = ±∞, b = ±∞ and NaN
        return false;

But it was causing the function to return false always, even for finite values. I think it should be:

if (!(std::isfinite(a) && std::isfinite(b)))  // a = ±∞, b = ±∞ and NaN
        return false;

[federico-busato]

thank you, @linomp!! I will fix it in the next release.