On Tue, 16 Sep 2014 13:29:57 -0700, Charles Mills wrote:

>Yup. Over-engineering. Over-validation. Like examining a filename to see if it 
>is syntactically valid rather than just passing it to the OS's open function 
>and seeing if it works for the OS.
> 
There's a partial and mostly unsatisfactory conjectural
explanation for the rule in what I see in the OUTPUT from
the FTP step:
 
EZA1736I   put //DD:SYSUT1     TEMP.FTPCAT                                      
EZA1701I >>> SITE VARrecfm LRECL=80 RECFM=V BLKSIZE=84                          
500 'SITE VARRECFM LRECL=80 RECFM=V BLKSIZE=84': command not understood. 

Apparently FTP attempts to inform the receiving system of the attributes
of the file by the content of the SITE command.  It tries to get these
attributes from the first catenand, making a couple bad assumptions:

o Attributes of that data set are valid iff it has been initialized.

o It has been initialized only iff non-empty.

o JES guarantees good attributes for an instream data set.

Also, there seems to be some evidence that FTP falters without
reporting qny error on a catenation of data sets with unlike
attributes.  Is this a documented restriction.

I hate MVS!  (z/OS?  OS/360?  Whatever?)

There's an aphorism to the effect that MVS is notorious for
solving (poorly, often, IMO) problems that would never arise
on any other OS.

-- gil

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

Reply via email to