Re: %decpos

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

Hi Rob,

You can't do that because %DEC() creates a variable.  Granted, it 
doesn't have a name, and you can't use it directly in your code -- but 
it's a variable, used by the compiler, under the covers.

That variable needs to be created at compile-time.  And, the %LEN BIF 
(and presumably also the %DECPOS BIF?) execute at run-time, and 
therefore the output of those BIFs can't be known at compile-time.  If 
it doesn't know the output at compile-time, it can't create a variable 
with those attributes at compile-time.

One workaround is to force the BIF to output to a named constant:

 D LEN             c                   %len(myDecimal)
 D DECPOS          c                   %decpos(myDecimal)
  /free
      mydecimal = %dec(rplary(x):LEN:DECPOS);


But, frankly, I never do that because I find it too cumbersome to code. 
  It's easier to just pick a really large output field, like:

      mydecimal = %dec(rplary(x):30:10);


Remember:  %DEC isn't controlling the size of the output to myDecimal. 
It's controlling the size of a temporary variable under-the-covers. 
Rplary(x) will be first converted to that temporary variable, and then 
separately copied to your myDecimal variable.

Think about that...   the size of the output of %DEC really doesn't matter.



On 7/24/2012 8:22 AM, rob@xxxxxxxxx wrote:
> Why can I do this:
> d mydecimal s 2p 0 inz(*zero)
> mydecimal=%dec(rplAry(x):%size(mydecimal):0);
> but not this
> mydecimal=%dec(rplAry(x):%size(mydecimal):%decpos(mydecimal));
> RNF0329: The third parameter for %DEC or %DECH is not valid.
>
> IBM Rational Developer for Power Systems Software
> 5724-Y99, 5733-RDG, 5733-RDH, 5733-RDJ, 5733-RDC
> Licensed Materials - Property of IBM
>
> Version: 8.0.3
>
> Rob Berendt