And, what if STATE would get a 3'th outputstream.  If defined: it would get
the invalid fileIDs, and STATE would no longer prematurely exit when
encountering an invalid fileID.

2007/8/29, Glenn Knickerbocker < [EMAIL PROTECTED]>:
>
> "John P. Hartmann" wrote:
> >  Let me understand your requirement.  Do you want a selection
> >  stage that will validate the record the same way as DMSVALDT does?
>
> That sounds just right.
>
> >  Note the STATE might reject file identifiers that < might be able to
> >  read (NAMEDEFS, essentially).
>
> I was actually thinking too narrowly, in terms of filemodes, and hadn't
> even considered unaccessed directories.  Accepting dirids but not
> namedefs, the same as STATE, sounds very reasonable.
>
> ¬R
>



-- 
Kris Buelens,
IBM Belgium, VM customer support

Reply via email to