🚀 go-pugleaf

In article <314db547.242588844@nntp.ix.netcom.com> miker3@ix.netcom.com (Mike Rubenstein) writes:

   mib@gnu.ai.mit.edu (Michael I. Bushnell, p/BSG) wrote:

   > You are not guaranteed that long is the widest integral type.  

   Huh?  You most certainly are guarantteed that long is the widest
   signed integral type and unsigned long is the widest unsigned integral
   type by ISO 6.1.2.5.  This with one caveat -- I am assuming that
   "integer type" and "integral type" have the same meaning, but I'm
   pretty sure I've read that this will be made clear in an upcoming
   technical corrigendum.

It appears you are correct.  Then the flaw is that you are prohibited
from using more than four widths of integers.  That's a mistake, which
we will all rue when the first 128 bit machines come out.  ;-)

Michael