On Thu, Sep 01, 2005 at 11:35:32PM +0200, MiKael wrote: > [EMAIL PROTECTED]:~$ rpncalc > rpncalc version 1.34. Copyright © 1993-2004 David Frey et. al. > Il s'agit d'un logiciel libre SANS AUCUNE GARANTIE. > Pour de plus amples informations, tapez 'warranty'. > Tapez 'quit' pour quitter ou '?' pour obtenir de l'aide. > 22 ! > 1: 1124000727777607680000 > 23 ! > 2: 1124000727777607680000 > 1: 25852016738884978212864 > > The result must be '25852016738884976640000'. > > I think it's a problem with the C double precision.
Yes. > So, it could be a good idea, if the fix is not easy to do, > to report this error to the user by an Overflow message for example. Yes. I should port the code from using doubles to arbitrary exact precision. I know that, but haven't had the time to implement it. Thanks, David
