Wrong result in (^)/2 bigint computation about trealla HOT 4 CLOSED

from trealla.

GNU Prolog doesn't have bigints. These Prolog systems
cannot realize (^)/2 correctly. You can try yourself the

ISO core standard Prolog flag "bounded", where
"true = no bigints" and "false = bigints":

?- current_prolog_flag(bounded, X).
X = true

Or alternatively you can try:

/* GNU Prolog 1.5.0 */
?- X is 51^20-1.
uncaught exception: error(evaluation_error(int_overflow),(is)/2)

from trealla.

Your result is probably a conversion back to integer
from float. To see that you need a precise float to
bigint conversion, which shows me that:

?- X is integer(1.1904242382761301E17).
X = 119042423827613008.

But the computation of a float is not that precise
as the computation of a bigint. Thats where the
precision is lost, and why the integer result shown

is not correct.

from trealla.

As already mentioned in another ticket, the error is gone.
For example it is now:

Trealla Prolog (c) Infradig 2020-2022, v2.7.15
?- X is 39^9, Y is 39*39*39*39*39*39*39*39*39.
   X = 208728361158759, Y = X.

Closing this ticket.

from trealla.