On 2016-12-02 10:58, Jesse 1 Robinson wrote:
> I'm a huge fan of both ISPF Stats and the Version/Mod values stored in PDS
> members columns 73 - 80. Preserving them requires a bit of forethought. When
> updating a member, it's common practice to create a new or backup member.
> Unfortunately, the 'simplest' ways are the worst:
>
Might PDSE member versions provide relief? Get your RFE in before
someone else wants it done wrong.
> -- ISPF supports Mod numbers from 00 - 99. There is no rollover, and the
> Version number is never incremented. This is a problem for a long-lived PDS
> like PARMLIB, which may last for years/decades. Eventually the Version
> reaches 99 and stays there despite subsequent edits. You can no longer
> compare versions by line number.
>
You might RFE (or RYO) a facility to refresh/compact mod numbers,
perhaps by dividing the content of 79-80 by 10 and adjusting
directory information.
> -- Remember that there are no ISPF stats for a PS dataset. Consider this fact
> when deciding what kind of file you want to maintain.
>
Interestingly, there are secret timestamps, visible only via NFS, on
PDSEs and their members. With µsec granularity! ("secret" as in
anyone who explained more to me would be obliged to kill me.)
These are maintained regardless of whether the library or member
was created by ISPF or by any other utility. It would be nice if
ISPF displayed those as a fallback when directory stats were nonexistent.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
