(34S) Integration (using the double-exponential method)

[emece67]

I'll change my program to return an error estimation in Y.

I've found that, after the change I made cause of Dieter's comment about epsilon, there's now a way to discern when a small result is due to roundoff when evaluating an integral that is zero or a small result due to an integral that is truly small, but different from 0. I'll try to add an heuristic to return 0 when the integral seems to be zero.

Regards.