class Float
def approx(other, relative_epsilon=Float::EPSILON, epsilon=Float::EPSILON)
difference = other - self
return true if difference.abs <= epsilon
relative_error = (difference / (self > other ? self : other)).abs
return relative_error <= relative_epsilon
end
end
100.2.approx(100.1 + 0.1) # => true
10e10.approx(10e10+1e-5) # => true
100.0.approx(100+1e-5) # => false
|