On Mon, Mar 4, 2013 at 1:57 PM, Gilles <gil...@harfang.homelinux.org> wrote:
> Hello. > > > >> I would like to release version 3.2 of [math] soon. >> Any thoughts about it? >> > > How soon? > There are many unresolved issues targetting 3.2. > There is at least one bug with no fix target (MATH-936). > I did look into this one already. The actual bug is quite simple to fix: * if 0.0 is returned by the RandomGenerator, return the lower bound The bigger problem is that the applied scaling is not robust: double scaled = r * upper + (1.0 - r) * lower + r; so for large values of lower/upper, as in the bug example this will lead to wrong results. Thomas