Re: 10^2 error. Bug or expected behavior?

[Barbara Morris <bmorris@xxxxxxxxxx>]

On 2025-06-18 11:18 a.m., Marco Facchinetti wrote:
> I guess %INT usually works but I found a note in a procedure we use to
> handle rounding that state: Use %INTH.
>
> Dcl-s intnbr    int(20);
>
> intnbr = %inth(xqta * (10 ** xnbrdec));

+1 to using %inth (or %dech) to get out of float. It's possible that the 
float value for 10 ** xnbrdec is say 99.99999999999999999999999 instead 
of 100, and %int would return 99.

It's usually better to get out of float as soon as possible, so I think 
it's better to do the %inth on the ** result.

intnbr = (xqta * %inth(10 ** xnbrdec));