Re: SFLNXTCHG head-scratcher

I do what you're trying to do all the time--it's one of  my standard 
subfile processing methods.

First, I read through all the changed records and validate the 
options/input, set the SFLNXTCHG on, and update the subfile record. Then I 
read through them and process them. But in-between, I reset the RRN 
(RRNSF1 in your case) to zero. If you don't, then you immediately get %EOF 
as there are no more records to process.

It doesn't require IO on the control record, or other finagling. You can 
use hidden fields and do your own comparisons, but why rewrite a basic OS 
function? Go down that path and you can end up writing a bunch of code 
that the OS already handles for you.

"RPG400-L" <rpg400-l-bounces@xxxxxxxxxxxx> wrote on 05/13/2015 12:56:25 
PM:
> ----- Message from Dan <dan27649@xxxxxxxxx> on Wed, 13 May 2015 11:
> 17:23 -0400 -----
> 
> To:
> 
> RPG400-L@xxxxxxxxxxxx
> 
> Subject:
> 
> SFLNXTCHG head-scratcher
> 
> It's been several years... meh, a decade since I last worked with 
subfiles
> prior to this job I started a few months ago.
> 
> Have a situation where SFLNXTCHG is not working as I expect it to.
> 
> DSPF:
>      A          R MAPTRA1S                  SFL
>      A  39                                  SFLNXTCHG
> 
> RPGLE:
>      FMAPTRADF  CF   E             WORKSTN INFDS(INFDS)
>      F                                     SFILE(MAPTRA1S:RRNSF1)
>      F                                     INDDS(DS@INDDS)
> ...
>      D DS@INDDS        DS
>      d SflNxtChg              39     39n
> ...
>       *** A field defined in the subfile record is updated by the 
program,
> then...
>      c                   Eval      SflNxtChg =
> *on                              SFLNXTCHG (for READC
>      C                   UPDATE    MAPTRA1S
>      c                   Eval      SflNxtChg =
> *off                             SFLNXTCHG reset
> 
> --->  At this point, should a READC pick up the record that was just
> updated?  In the program I'm modifying, the next I/O operation on the
> workstation file is a READC on the subfile.  Does there need to be an
> intervening EXFMT / WRITE of the subfile control record?  I tried a "1
> SETLL  MAPTRA1S" after the update above, but the compiler isn't buying 
it.
> (Also tried *LOVAL in factor 1.) Here is the code with the EOF test:
> 
>      C                   READC(E)  UGLTRA1S
>      C                   DOW       %EOF = *OFF
> 
> Given the earlier code, I would have expected the READC to find the 
updated
> record, but debug is showing that %EOF is *on, and is skipping the DOW
> group.
> 
> Ideas, advice appreciated!
> 
> - Dan