For concatenating dataset overrides not being ironic. I gave an example the 
first DD of a concatenation determining the attributes of the open. 
Concatenation is not a requirement but in this instance, it helped because it 
allowed AMATERSE to obtain the attributes it needs.

I admit I don't re-read JCL and I mostly use existing JCL. But that doesn't 
make my statement false. For LRECL, it clearly states "LRECL applies to data 
sets with the BPAM, BSAM, EXCP, QISAM, QSAM, and TCAM access methods, and with 
SMS, to VSAM data sets.". Just because UNIX files are implemented thru one of 
these, doesn't mean UNIX files are that access method. It later states in 
OVERRIDES "LRECL overrides the record length specified in the data set label". 
UNIX files don't have a data set label or lrecl. IBM modified IEBGENER (and 
other utilities) to support DCB for UNIX files so yes they do work. I brought 
up IEBGENER with regards to VSAM to show that not all dataset types are 
supported. I did not say anything about IEBGENER supporting UNIX files.

As for mentioning GIMZIP and ZIP, I was proposing them as alternatives to 
AMATERSE. Since there wasn't any context, I made a false assumption.

Jon Perryman.



>________________________________
> From: Paul Gilmartin <[email protected]>
>To: [email protected] 
>Sent: Friday, November 8, 2013 8:58 AM
>Subject: Re: Aging Sysprogs = Aging Farmers
> 
>
>On Thu, 7 Nov 2013 22:36:30 -0800, Jon Perryman wrote:
>>
>>As for concatenating an MVS dataset before the UNIX file isn't as ironic as 
>>it seem's. This actually occurs more often than you think. The classic 
>>example is specifying a block size on the first dataset that is larger than 
>>it's block size.
>>
>Perhaps by programmers who don't understand that it suffices to
>specify an overriding BLKSIZE on the DD statement.  Or is there
>some context where the concatenation is necessary.
>
>>That is exactly what is happening with the UNIX file. UNIX files don't 
>>support an
>>RECFM, LRECL or BLKSIZE so specifying them on the DD results in them being 
>>ignored. 
>>
>Utterly false!  Have you never tried it, nor read the JCL Reference?  I do
>it routinely, and it works very well on, for example on SYSUT1 for IEBGENER;
>no concatenation required.
>
>>When the first DD specifies a dataset that supports RECFM, LRECL and BLKSIZE,
>>AMATERSE now has information it needs. Secondary DD's are not scrutinized as 
>>much as the first dataset in a concatenation.
>>
>Fortunately, it's better than it used to be.  Long ago, the BLKSIZE of the 
>first
>catenand dominated; nowadays the maximum prevails.
>
>>Maybe GIMZIP (from SMP/E) or zip from the UNIX utilities page might help.
>>
>I doubt that either of these is capable of understanding an AMATERSE
>archive, but I haven't tried it, so here I'm speaking from ignorance.
>
>-- gil
>
>----------------------------------------------------------------------
>For IBM-MAIN subscribe / signoff / archive access instructions,
>send email to [email protected] with the message: INFO IBM-MAIN
>
>
>

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

Reply via email to