isfinite(x) misreports float values <= 0 as not finite.

[d-leinhaeuser]

glm::isfinite from compatibility.inl returns false for float or double values that are less than or equal to 0.

The bug is in line 49:

return x >= std::numeric_limits<genType>::min() && x <= std::numeric_limits<genType>::max();

For floating point types with denormalization (see std::numeric_limits::has_denorm) std::numeric_limits::min() returns the smallest finite number greater than 0 that the type can represent instead of the smallest finite number overall.

The following should give the correct result for integer and floating point types:

if (std::numeric_limits<genType>::is_integer || std::denorm_absent == std::numeric_limits<genType>::has_denorm)
{
    return std::numeric_limits<genType>::min() <= x && std::numeric_limits<genType>::max() >= x;
}
else
{
    return -std::numeric_limits<genType>::max() <= x && std::numeric_limits<genType>::max() >= x;
}

[Groovounet]

This issue is fixed in master branch for GLM 0.9.7 release

Thanks for contributing!