On Mon, Apr 18, 2011 at 06:02:11PM +0200, Matthias Drochner wrote: > > There should be no rounding involved -- the only case I can > imagine where this matters is +0.0 vs -0.0. I see - thanks! Martin