On Mon, Dec 09, 2013 at 08:40:38AM -0500, Pavel Simerda wrote: > There's no reason for negating the error values in the current code. > looks ok for me, I wonder how negating errors originated.