Re: floats to integers?

[hirasuna@acm.org (Michael Hirasuna)]

In article <EACHUS.95Feb14104008@spectre.mitre.org>,
eachus@spectre.mitre.org (Robert I. Eachus) wrote:

>    Hmmm... If you want to always return a valid answer when possible,
> you need to do:
> 
>     function Floor (X: in Float) return Integer is
>       Temp: Integer;
>     begin
>       Temp := Integer(X);
>       if Float(Temp) > X then return Temp - 1; else return Temp; end if;
>     exception
>       when others =>
>         if X < Float(Integer'LAST) + 1.0 and X >= Float(Integer'LAST)
>         then return Integer'LAST;
>         else raise Constraint_Error;
>         end if;
>     end Floor;

I receive another solution from Angel Alvarez (aalvarez@goya.eunet.es)

if Float(Integer(X)) = X 
  then return Integer(X);
  else return Integer(X-0.5);end if;

At first glance this seems a more brute force approach, but is does avoid
the problems of the possible numeric error that may arise from integer
substraction.

-- 
Michael Hirasuna    |  hirasuna@acm.org