On 2016-09-06 12:39, Richard Pinion wrote:
> I was surprised when it happened to me.  I made
> the mistake of assuming it would work as other 
> programs do, COBOL programs and non-OMVS utilities.
> 
> However, when I went to the manual it explicitly stated that
> was the way it worked.  So you're probably right. 
>
RFE?  FTP ought to behave as much as possible like QSAM processing
the same input.  Principle of least astonishment.  And a good
example of how extensive documentation is not an alternative
preferable to a quality product.

Hmmm... Tape convention?  Two consecutive tape marks?  No excuse.

Likewise, FTP seems to ignore overriding attributes in a DD statement
and lets the F1 DSCB dominate.

I wonder what it does with a concatenation of Classic and UNIX files.

> And TRSMAIN will be quite happy to explode an
> incomplete input data set, giving a zero RC.
> 
SR?  I'd call this a data integrity failure.  But perhaps the
TERSE format was specified with insufficient attention to
redundancy and error detection.

> -----Original Message-----
> From: Richard Pinion
> Sent: Tuesday, September 06, 2016 10:57 AM
> 
> And one thing to be aware of when using concatenated data sets as input to 
> z/OS FTP.  If the first or an intervening data set is empty (assuming a valid 
> EOF has been written to the empty data set) and the following data sets are 
> not empty, EOF is signaled without processing the following non-empty data 
> sats.

-- gil

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

Reply via email to