In a recent note, "Shmuel Metz (Seymour J.)" said:
> Date: Sat, 28 Apr 2007 22:31:43 -0300
>
> In <[EMAIL PROTECTED]>, on 04/26/2007
> at 02:22 PM, Paul Gilmartin <[EMAIL PROTECTED]> said:
>
> >So, now I must RTFM "Using Data Sets" to find the section that states
> >that JCL concatenation of HFS files _is_ supported, since OS/390?
> >Earlier? So I can start YA PMR.
>
> Not unless everything in the SYSLIN concatenation is an HFS. The
> application determines whether and how concatenation of unlike
> attributes is supported. But this sounds like a good candidate for
> submitting a requirement, if you can come up with a business case.
>
Irrelevant; in this case I did not have unlike attributes.
> >Why can't IBM's developers learn IBM's own rules?
>
> What rules are they not following?
>
Title: z/OS V1R8.0 DFSMS Using Data Sets
Document Number: SC26-7410-06
3.9.7 Concatenating UNIX Files and Directories
3.9.7.1 Sequential Concatenation
To process sequentially concatenated data sets and UNIX files, use a DCB
that has
DSORG=PS. Each DD statement can specify any of the following types of data
sets:
* Sequential data sets, which can be on disk, tape, instream (SYSIN),
TSO/E terminal,
card reader, and subsystem (SUBSYS)
* UNIX files
* PDS members
* PDSE members
The text is ambiguous whether two or more of the four bulleted types can be
concatenated (e.g. instream with UNIX or PDS members with PDSE members).
But that seems implicit from the section title. And, empirically it works
with applications other than Binder.
> >And why must they go out of their way to write code to enforce the
> >rules they invent in their delusions?
>
> What is your evidence that they did in this case? It sounds to me like
> something that you've asked for in the past; an intelligible message
> instead of, e.g., an ABEND.
>
On the contrary, I have generally opposed this sort of prior censorship;
the application should attempt the OPEN and report the result rather
than attempting to predict (incorrectly in this case) success or failure.
> >Why not just read SYSLIN with QSAM and let it work or fail as it
> >may.
>
> Because then you'd whine about the ABEND or other error.
>
Only if it were to fail.
> >If they don't deliberately try to make it fail, it will work.
>
> How?
>
Well enough for me. I have regularly concatenated e.g. instream data
sets with UNIX files with applications other than Binder (IEBGENER,
SMP/E, ...) and observed the desired and expected success. It appears
that the DFSMS group has done its job well here; the Binder group
simply has a deficiency of faith in DFSMS.
-- gil
--
StorageTek
INFORMATION made POWERFUL
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
