On Thu, Apr 26, 2018 at 11:50 AM, Paul Gilmartin <
[email protected]> wrote:

> On Thu, 26 Apr 2018 07:33:39 +0200, Peter Hunkeler  wrote:
> >
> >>IMHO it is problem with IBM. Are you deaf?
> >Sometimes it is; sometimes they are. However, not this time.
> >
> >>People really NEED the functionality to simply make single dataset
> archive, transmit it to PC *with no tricks and black magic* and then upload
> it to a z/OS and unpack.
> >
> >Know your job. What is difficult about running two programs one after the
> other to get the desired result?
> >
> IBM could make it easier, rather than assuming the burden of providing
> full employment for customers' systems programmers.  Ideally:
>
> All archiving facilities such as ADRDSSU, TERSE, TSO TRANSMIT, pax,
> IEBCOPY,
> ... should treat their archives as featureless streams which can be wrtteh
> to CKD DASD, tape, directed to UNIX files (which could be piped to network
> or transmitted via FTP BINARY), ...  Extraction should ignore record
> boundaries
> in those archives.
>

​I agree. Given that XMIT only works with FB/80, I think it would be
"reasonable" for it to "assume" that any UNIX data stream fed to it is
implicitly FB/80 and just process incoming data in 80 byte "chunks" (aka
records).​



>
> pax does this well.  Its archives can be CKD data sets (and it's
> nondiscriminatory
> concerning DCB attribute) or UNX files, and I can usefully:
>     pax -w  directory | ssh some.host "pax -vr"  # (Simplified.)
>
> Example:  GIMZIP relative files are IEBCOPY PDSU which GIMZIP transforms
> to UNIX
> files and RECEIVE FROMNTS transforms back to Classic PDSU.  Considerable
> overhead
> in processing and allocation of temp ds could be avoided if IEBCOPY deaht
> directly
> with UNIX files as PDSU.
>

​Unfortunately, IEBCOPY already has the "cruft" of using physical size of
the block written to disk as implicit data describing the schema(?). What I
have wondered about is possible ability of "emulating" RECFM=U by using a
UNIX file with FILEDATA=RECORD. The "metadata" which is encoded by the
physical size of the block would be replaced by the explicit 4 byte "record
length" inserted by the access method due to the FILEDATA=RECORD. ​



>
> ... the others shoulc take a c[l]ue.
>
> -- gil
>
>

-- 
We all have skeletons in our closet.
Mine are so old, they have osteoporosis.

Maranatha! <><
John McKown

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to