Re: Apparent bug-fix in RPG compiler breaks program

Thanks Barbara!

You hit it on the head, the subfields were defined in a file.

Unqualifying the DS is the solution we went with.

Charles

On Jun 27, 2017 18:34, "Barbara Morris" <bmorris@xxxxxxxxxx> wrote:

On 2017-06-27 1:59 PM, Charles Wilt wrote:

...

but in the calculations, the code has...
lvData.lvFees(lvCnt) <snip>

Compiling on a v7r2 box, with TGTRLS(*CURRENT), the compile fails
with RNF7591: An operand of a qualified name expression is not valid.
...

We added the QUALIFIED keyword to the DS, and the compile succeeds. But
the program doesn't work correctly.

Check the cross reference to see if the names $FEE01, $FEE02, $AMTTYP01 etc
exist as standalone fields in the program. If so, check whether there is
any code that is working with the unqualified fields, or if there are
compiler-generated I specs that load those fields from a file. If so, then
any code that uses those standalone fields used to work with the subfields
of the data structure, and now they're just working with the standalone
fields.

If that's what's happening, then instead of adding the QUALIFIED keyword,
you should change lvData.lvFees(lvCnt) to just lvFees(lvCnt).

Or, if you want to keep the data structure qualified, then if $FEE01 etc
are coming from a file, you could add keyword PREFIX('LVDATA.') to your
file, and then the fields would become subfields of the qualified data
structure.