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.

Charles

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Paul Gilmartin
Sent: Tuesday, September 16, 2014 1:21 PM
To: [email protected]
Subject: Re: z/OS FTP client behavior

On Tue, 16 Sep 2014 11:40:58 -0700, Richard Pinion wrote:
>
>To prevent transferring data from an empty file, FTP checks whether the first 
>file in a concatenation series is empty and allocates an empty data set. No 
>data is transferred.
>
Morons.  Obsessive-compulsive morons.  The test is otiose.  If any catenand is 
empty, EODAD will immediately be entered for that catenand and no data will be 
transferred.  But that shouldn't prevent transferring the remaining catenands.  
There is no more reason for FTP to treat "empty" as a special case than for 
access methods to do so in general.

BTW, what happens if the first is non-empty, but the second is empty?
Are third and subsequent catenands transferred.

This strikes me as comparable to the JCL behavior that "PATH='/dev/null'"
is replaced by "DUMMY".  Probably an ancient MVS hand with tunnel vision found 
a statement in a Users Guide that "/dev/null is like DUMMY" and opened an SR 
that catenands after /dev/null were nonetheless processed.  Support agreed with 
his ignorance and made the change.

And once such conceptual blunders are codified in the documentation they become 
part of the design, which can never be ameliorated.

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

Reply via email to