On 2017-03-17, at 10:32, Jesse 1 Robinson wrote: > Today's posts moved me to actually look at my GLOBAL DDDEFs. My SMPPTS% > entries are defined discreetly, not as a concatenation of libraries. So I > have > > SMPPTS > SMPPTS1 > SMPPTS2 > OK. Slip of the mind #1. I was thinking of SMPNTS, not SMPPTS.
> pointing to individual libraries. I believe this is what causes SMP/E to > examine a succession of SMPPTS% DDDEFs until a sysmod is found or the next > n+1 entry does not exist. There is no place to specify something like > 'highest PTS DDDEF defined'. > The simple way to do this would be to perform the search in the order in which the DDDEFs appear in the zone definition, rather than in numeric order. > All this is to say that there is no actual SMPPTS concatenation, rather a set > of libraries that will be searched in order until 'the next one' is not > found. That's why you can freely reallocate the set of PTS libraries and move > sysmods around without disrupting SMP/E. Also why you cannot skip any SMPPTS% > in the series, which would terminate the search. > > Could SMPPTS be defined as an actual concatenation of libraries? That's > conceivable within the framework of SMP/E, but I don't know if it would work > in practice. A perfect task to assign to your favorite sysprog intern. Ha ha. > > Slip of the mind #2. I used to know that you can't write to a concatenation, neither partitioned nor sequential. I keep misleading myself, whether by ISPF's list of library specifications or by CMS's read-only extensions to drive letters. I'll assume that RECEIVE checks the entire (consecutive; gaps ought to be reported as errors) SMPPTS% list to verify that the REWORK number is higher. If it is, it RECEIVEs the SYSMOD. Does it then scratch the older instance, even if it appears later in the SMPPTS% list? Apologies, gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
