Hi, On Tue, 9 Aug 2011, Richard Guenther wrote:
> > Actually we should adjust the ranges only if it's either all positive > > or negative values from the start and the adjustment keeps us on the > > same side of zero. > > So we kept discussing this to death somewhat. Which eventually resulted > in me doing some exhaustive testing and thinking again - which makes me > end up with the following, which passed bootstrap on > x86_64-unknown-linux-gnu and is now in testing. Sigh. So the strange asymmetry with the [0,1] & [-2] example that itched me really pointed to a problem, we just couldn't pinpoint it. We'll have to upgrade our white-board-calculating skills ;) Ciao, Michael.
