Thanks for the answer. I actually assumed that f>d puts exact 64 bit 
representation of a floating on the floating stack to the integer stack - 
nondestructively, even though it starts looking as an integer on the integer 
stack. Too sad it is not so. 

This raises another question. Can anyone tell me a trick how to save and then 
restore floating stack numbers ( 1 or 2 ) for doing intermediate calculations 
which require all floating stack slots? When I do computations then saving 
iteration to an array in memory is fine. But for a sum when at each iteration 
I need to increment the same value fetching from/to memory seems costly 
(maybe I am wrong?)

Thank you!

On Sunday 11 June 2006 04:33, Bernd Paysan wrote:
> On Sunday 11 June 2006 10:45, Sergey Plis wrote:
> > Hi!
> >
> > It looks like there is one more bug. Looking at the session below you
> > see that during conversion using f>d digits after decimal point are
> > lost. Is this expected semantics or a bug?
>
> This is expected semantics - d is an integer.
>
> > order FLOAT FORTH FORTH ROOT   FORTH  ok
> > !2 f>d  ok
> > .s <2> 2 0  ok
> > d>f f. 2.  ok
> > !0.2 f>d  ok
> > .s <2> 0 0  ok
> > d>f  f. 0.  ok
> > !1.2 f>d  ok
> > .s <2> 1 0  ok
> > d>f f. 1.  ok
> >
> > Thank you!

-- 
Sergey Plis
PhD Candidate
University of New Mexico

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to