Re: Odd problem passing a data structure to a procedure

["Rory Hewitt" <roryhewitt@xxxxxxxxx>]

David,

Check out this thread on the systemiNetwork:
http://www.systeminetwork.com/isnetforums/showthread.php?t=45802. Although
it deals with the specific issue of duplicate CONST parameters, it does
include some code to allow you to get any OpDesc information without using
CEEDOD...

Now I don't know whether the problem with CEEDOD returning a 0 for a DS is
with CEEDOD or is with the fact that the OD is passed incorrectly, but if
it's the former, you could try using the following code in a copybook to
bypass uisng CEEDOD:

D NPMPARMLISTADDR...
D                 PR              *   Extproc('_NPMPARMLISTADDR')
*
D NPMPARMLIST     DS                  Based(NPMPARMLIST@) Qualified
D   OpDescList@                   *
D                               16A
D   Argument@                     *   Dim(400)

D OpDescListDS    DS                  Based(OpDescListDS@) Qualified
D   NbrOfParms                  10I 0
D   Reserved1                    8A
D   ArgOpDesc                   10I 0
D   Reserved2                   16A
D   OpDesc@                       *   Dim(400) *
D OpDesc          DS                  qualified based(OpDesc@)
D   desc_type                    3I 0
D   data_type                    3I 0
D   desc_inf1                    3I 0
D   desc_inf2                    3I 0
D   data_len                    10I 0

and in your program use the following code instead of calling CEEDOD:

NPMPARMLIST@ = NPMPARMLISTADDR();
if NPMPARMLIST@ <> *null;
 OpDescListDS@ = NPMPARMLIST.OpDescList@;
 if OpDescListDS.NbrOfParms > 0;
   for ParmIndex = 1 to OpDescListDS.NbrOfParms;
     if NPMPARMLIST.Argument@( ParmIndex ) <> *null;
       OpDesc@ = OpDescListDS.OpDesc@( ParmIndex );
     endif;
   endfor;
 endif;
endif;

Try it out and let us know if it works.

Rory


On 12/1/06, David Gibbs <david@xxxxxxxxxxxx> wrote:
> Lim Hock-Chai wrote:
> > CEEDOD will return 0 if caller passes a data structure.  This problem
> > has been fixed on V5R3.
>
> Yep, found that.
>
> > Lets have a vote:  :)
> > Should IBM creates PTF to make the above fix available on previous OS?
>
> Well, you have my vote ... I'd be happy to have it retrofitted to V5R2
> (as that's the release that I'll be able to target pretty soon).
>
> david
> (who wonders if Barbara Morris can be bribed with a midrange.com button?)