Re: ovrdbf scope

[Scott Klement <midrange-l@xxxxxxxxxxxxxxxx>]

Joel,

By default, an open data path (ODP) is _not_ shared across two programs. 
 This can be chagned by specifying SHARE(*YES) on the CRTPF/CRTLF 
commands, or on an override: OVRDBF SHARE(*YES)

(But, your override doesn't seem to be doing that... at least, not in 
the code you posted.)

So, by default, PGMB can't interfere with PGMA's cursor position.  But, 
if someone has changed the file (either via CRTPF/CHGPF/CRTLF/CHGLF or 
OVRDBF) to SHARE(*YES), this is causing your grief.

I've seen this many times in the past in code that was rewritten from 
S/36 OCL.  The OCL/36 programmer would see
// FILE NAME-MYFILE,DISP-SHR in the OCL, and think they needed an 
equivalent to DISP-SHR in CL, so they'd code SHARE(*YES). This was just 
confusion on their part -- they do totally different things.

-SK


On 5/22/2012 2:17 PM, Stone, Joel wrote:
> I would like to
>
> OVRDBF     FILE1  position(*KEYAE 1 *N&File1Key)
>
> Loop1:  RCVF
>
> Process record ...
>
> Call pgmB  (which happens to read and reposition FILE1 cursor)
>
> If not EOF goto loop1
>
> When the pgm loops back to Loop1, the cursor has been repositioned by
> pgmB, which messes up my loop.
>
> Is there anything I can do with the top OVRDBF to ensure that it
> ignores any repositioning of the file cursor in programs down the
> stack?
>
> Or must I save off the key values and reposition the file with
> another ovrdbf?
>
>
> Thanks!