On Fri, Aug 26, 2016 at 09:36:51AM +0200, Martin Husemann wrote: > > IBM has its own float format with FLT_RADIX=16. > > Only in 1998 they started offering IEEE754 float on the same hardware. > > Does NetBSD intend/does support this? is this something to keep in mind? > > I'd say: worry about this when a port to the relevant hardware comes up. > > The VAX pre-IEEE754 format is enough to deal with already, and will likely > provide a good guide when adding other non-754 formats later (if ever).
It matters because of whether scalbn == ldexp... (that was how this came up) -- David A. Holland dholl...@netbsd.org
