Re: C_FLOAT New user question

[Keith Hopper]

In article <address@hidden>,
   Norbert Nemec <address@hidden> wrote:
> AFAIK, Sather tries to be IEEE... correct concerning floats. Strange 
> that the conversion does not work.

> On Mon, Oct 02, 2000 at 05:07:16PM +0000, Dave Simons wrote:


> >             loop
> >                     USING_SATHER::show_float(#(number));
> >                     USING_C::show_float(#(number));
> >                     number := number + 0.2;
> >                     until!(number > 3.0);
> >             end;
> >     end;
> > end;
> > 
> > class USING_SATHER is
> >     show_float(sather_number:C_FLOAT) is
> >             #OUT+"Number according to Sather: "+sather_number.flt+"\n";
> >     end;
> > end;
> > 
> > external C class USING_C is
> >     show_float(c_number:C_FLOAT);
> > end;

Have you thought of a big/little endian problem.  This is one of the things
which we discovered in the compiler - it is the order in which the two
32-bit bit-patterns get stuffed onto the stack - but then its only
approximate arithmetic after all!!

              kh

-- 
City Desk
Waikato University
[PGP key available if desired]