Experimental Physics and
Industrial Control System

"Schoeneburg, Bernd" <[email protected]> · Mon, 19 Oct 2009 11:38:45 +0200

Hi all,

I found that the get_precision function is not correct in the following

records:

- calc
- calcout
- dfanout

My change in the calc record (3.14.10) (similar for others):
163,167c163,167
<     if (paddr->pfield == (void *)&pcalc->val) {
<       *pprecision = pcalc->prec;
<     } else {
<       recGblGetPrec(paddr, pprecision);
<     }
---
>     *pprecision = pcalc->prec;
>     if (paddr->pfield == (void *)&pcalc->val)
>       return 0;
>
>     recGblGetPrec(paddr, pprecision);

The function recGblGetPrec does only limit the precision to [0,15] if

the type is float or double. Setting is done only for integer types. The

result of the old code was a zero-precision display of non-VAL fields.

This change is necessary for Epics 3.14.11, too.

- Bernd

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System