rlib-users

[Rlib-users] RLIB 1.3.3: function abs() works only for integers

[Derek G. <dgi...@ge...>]

RLIB Gurus--

The RLIB function abs() works only for integers, which explains why =
float values passed to this function break in our reports. I refer you =
to pcode_op_functions.c lines 688 through 701:

gint rlib_pcode_operator_abs(rlib *r, struct rlib_value_stack *vs, =
struct rlib_value *this_field_value) {
    689         struct rlib_value *v1, rval_rtn;
    690         v1 =3D rlib_value_stack_pop(vs);
    691         if(RLIB_VALUE_IS_NUMBER(v1)) {
    692                 gint64 result =3D =
abs(RLIB_VALUE_GET_AS_NUMBER(v1));
    693                 rlib_value_free(v1);
    694                 rlib_value_stack_push(vs, =
rlib_value_new_number(&rval_rtn, result));
    695                 return TRUE;
    696         }
    697         rlib_pcode_operator_fatal_execption("abs", 1, v1, NULL, =
NULL);
    698         rlib_value_free(v1);
    699         rlib_value_stack_push(vs, =
rlib_value_new_error(&rval_rtn));
    700         return FALSE;
    701 }

It calls only the standard abs() function and doesn't attempt to check =
if the number is a double so that it could instead call the standard =
fabs() function.

I could take a crack at fixing this if you'd prefer, but I wanted to =
bring it to your attention.

Cheers

--
Derek Giromini
Gelber Group, L.L.C.
desk: +1.312.692.2843
pager: +1.312.689.0378

Re: [Rlib-users] RLIB 1.3.3: function abs() works only for integers

[Bob D. <bd...@si...>]

Derek,

Are you sure?

Try: 

<field format="'%.5d'" value="abs(-5.3333)" width="10"/>

It works for me.

Also RLIB doesn't need to call fabs because the number is stored in FXP
and not floating point notation for numerical accuracy

- bob


On Thu, 2005-05-12 at 13:34 -0500, Derek Giromini wrote:
> RLIB Gurus--
> 
> The RLIB function abs() works only for integers, which explains why
> float values passed to this function break in our reports. I refer you
> to pcode_op_functions.c lines 688 through 701:
> 
> gint rlib_pcode_operator_abs(rlib *r, struct rlib_value_stack *vs,
> struct rlib_value *this_field_value) { 
>     689         struct rlib_value *v1, rval_rtn; 
>     690         v1 = rlib_value_stack_pop(vs); 
>     691         if(RLIB_VALUE_IS_NUMBER(v1)) { 
>     692                 gint64 result =
> abs(RLIB_VALUE_GET_AS_NUMBER(v1)); 
>     693                 rlib_value_free(v1); 
>     694                 rlib_value_stack_push(vs,
> rlib_value_new_number(&rval_rtn, result)); 
>     695                 return TRUE; 
>     696         } 
>     697         rlib_pcode_operator_fatal_execption("abs", 1, v1,
> NULL, NULL); 
>     698         rlib_value_free(v1); 
>     699         rlib_value_stack_push(vs,
> rlib_value_new_error(&rval_rtn)); 
>     700         return FALSE; 
>     701 }
> 
> It calls only the standard abs() function and doesn't attempt to check
> if the number is a double so that it could instead call the standard
> fabs() function.
> 
> I could take a crack at fixing this if you'd prefer, but I wanted to
> bring it to your attention.
> 
> Cheers
> 
> -- 
> Derek Giromini 
> Gelber Group, L.L.C. 
> desk: +1.312.692.2843 
> pager: +1.312.689.0378
>