fix bug in time diff on unix

[davidhodo]

When comparing times if the second changes, that change should be taken into account calculating the change in nanoseconds. If not you can end up with negative differences when the second field increments.

[wjwwood]

Well, I believe you that there might be a problem, but according to this documentation:

http://www.gnu.org/software/libc/manual/html_node/Elapsed-Time.html

...that code is correct already.

But I think the bug is that the nanosecond diff value can be negative, i.e. the nano seconds of the end might be smaller than the nano seconds of the start even though end comes after start in total time... in which case it should be diff.useconds = (1e9 - abs(end.nanoseconds - start.nanoseconds) / 1000 and diff.seconds -= 1.

Can you try that change instead and I'll also play around with it.

[wjwwood]

Checkout #28 and let me know if that works for you.