%DECPOS doesn't work with %DEC

["Buck" <buck.calabro@xxxxxxxxxxxx>]

I am preparing a class on RPG IV for internal use, and was tinkering with
the examples in the RPG Reference regarding intermediate precision.
Straight from the reference, I have this:

d fld1            s             15p 4
d fld2            s             15p 2 inz(4)
d fld3            s              5p 2 inz(100)
d fld4            s              9p 4 inz(1)
d fld5            s              9p 4 inz(2)

eval      fld1=fld2/(((fld3/100)*fld4)+fld5)

Of course, it returns 1.0000, exactly as expected.  In an attempt to
demonstrate (again, from the manual) that %DEC can be used to palliate this,
I tried:

eval

            fld2/(((fld3/100)*fld4)+fld5):
            %size(fld1):
            %decpos(fld1))

*RNF0329 30 The third parameter for %DEC or %DECH is not valid.

Ah, well.  That's too bad, because it would open a programmatic way to
influence the intermediate calculations in the same way that re-ordering
Traditional Operation Codes did - in essence, twisting the compiler's arm to
use the intermediate results that *I* want.  As it stands, I need to do
this:

eval

            fld2/(((fld3/100)*fld4)+fld5):
            %size(fld1):
            4)

Just an observation for anyone interested in intermediate precision, and
detecting its results programmatically.
  --buck