Re: simple problem with string handling

Different folks, different strokes - comparing DS' is often recommended, even in these lists. This behavior makes that not so useful.

I'm not advocating it necessarily, I'm advocating asking IBM to look at it. The RFE process is a very good way to do this, as is the COMMON requirement process. Hey, I might even submit a requirement directly, since I'm on the CAAC. And someone else can submit a DCR.

I was told long ago, it never hurts to ask - the worst you can get is a NO, and then nothing has changed, anyhow.

Cheers
Vern

On 2/23/2013 10:52 AM, John Yeung wrote:

On Sat, Feb 23, 2013 at 11:17 AM, Vernon Hamberg
<vhamberg@xxxxxxxxxxxxxxx> wrote:

I understand that the processing (stand-alone) varying fields is not a
bug - my post probably didn't say that clearly enough, although I tried
a little. At the same time, there's no reason not to report it as such
in the context of a data structure subfield, in order to see if
something can be done. Perhaps the treatment of varying could be
adjusted when in a data structure, because of the side effect described.

I don't believe anything *should* be done. The point of a DS is to
just be a contiguous chunk of memory. You could have a 10-byte DS
that has a thousand overlapping subfields. The proposal to treat
varying subfields differently is tantamount to proposing that
comparison of two such data structures is no longer just comparing 10
bytes to 10 bytes but rather 1000 fields to 1000 fields.

For good or ill, a data structure in RPG is just a byte buffer.

John