Re: Renaming Fields in the D specs

[boldt@xxxxxxxxxx]

Bob wrote:
>The example that Mark sent is close to correct, he omitted the "E" in
>column 22 (I think?) so the example should be like this:
>
>> D xternalDS  E DS                           extName(FileName)
>> D  field1    E                              extFLD(newField1)
>> D   field2   E                              extFLD(newField2)
>
>FYI, this is #3 dumb syntax in RPG IV on my list of top ten dumb syntax
>issues in RPG IV. Why do we need an "E" if the EXTFLD() keyword is
>there?

Bob:  I agree the syntax isn't pretty, but there were reasons
for the design, and I'm sure you were told the reasons a long
time ago.

For the benefit of others, the justification for the design is
rooted in the need for RPG IV to be functionally compatible
with RPG III, which allowed program described subfields to be
added to the end of externally described data structures.  The
E in position 22 was put there to distinguish the program
described subfields from externally described subfields.

Goofy syntax?  Sure, OK.  But that's not the worst problem with
externally described data structures.  The bigger problem is
that EDS's only include the input-capable fields, which makes
EDS's pretty much useless for WORKSTN record formats and output
files.  More goofy still is that it's not just input-capable
fields, it's the fields that are listed by the API QUSLFLD as
having input buffer end positions, which for some reason
includes some output-only fields in certain situations!

Anyways, I didn't post to get into a debate with Bob Cozzi on
RPG IV language design.  I just thought I'd use the opportunity
to note that this is one area where RPG IV programmers should
expect some improvements in a future release.  People have
asked for some more flexibility with ext-desc I/O, and allowing
DS names on ext-desc I/O operations is one thing we're looking
at.  This would, of course, require improvements to EDS's.
(You should also expect some improvements in the way search
arguments are specified for keyed I/O operations.)

Cheers!  Hans







Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com