Re: gp2c question

[John Cremona <john.cremona@gmail.com>]

Thanks for the quick reply, Bill!

I'm not worried about the warning, as long as nothing actually bad is
happening.  I had tried changing the constants 1 and 2 in the
expression to 1.0 and 2.0, but there were still these warnings.

John

On Thu, 10 Sep 2020 at 14:42, Bill Allombert
<Bill.Allombert@math.u-bordeaux.fr> wrote:
>
> On Thu, Sep 10, 2020 at 02:05:55PM +0100, John Cremona wrote:
> > I have a gp function which I am running using gp2c.  In the function I
> > have the lines
> >
> > p = 20.0;
> > a = random(p)*2 -1;
> >
> > since I want a to be a random real in the balanced unit internal
> > [-1,1].  (I hope that is correct gp:  ?random does not mention the
> > possibility of an argument of type t_REAL, but ??random does.)
> >
> > Both p and are declared in a my() statement as type real with
> > my(p:real, a:real, ...). The second line above causes a warning from
> > gp2c
> >
> > Warning:MCq.gp:9: Assignement to a less precise type: real<-gen
> > a=(random(p)*2)-1
> >
> > Should I worry about that?    And surely the right and left hand sides
> > have the same type anyway?
>
> The issue is that gp2c currently does not know that random(real) is a
> real, and it does not know that real*2 is a real either, because it uses
> a rule 'real*int' and real*0 is not a real.
>
> You can shut up the warning by adding an explicit cast:
> a=((random(p)*2)-1):real
>
> I will add the necessary data to PARI to shut up the warning.
>
> Cheers,
> Bill
>