On Wed, 17 Dec 2025, at 19:26, Paul Gilmartin wrote:
> On Wed, 17 Dec 2025 18:36:00 +0000, Jeremy Nicoll wrote:
>> ...
>>Address ISPexec
>>"lminit dataid(did#i) dataset('"file#i"')"
>>"lminit dataid(did#o) dataset('"file#o"')"
>>
> Does that actually allocate those two data sets? Or ...
Yes - & you can specify the enqueue type.
I mostly used the ddname approach - I think because
allocation failures were easier to handle before trying
to use ispf services.
>> ...
>>"lmcopy fromid("did#i") todataid("did#o")" ,
>> "frommem(a*) replace"
>>say 'lmcopy rc:' rc
>>
> Does that perform then free both allocations? Then ...
I don't think it would free anything; once you've got
the dataids returned by ispf you could make lots of
"LMxxx" calls so you'd expect the dataids still to
work, so the underlying allocations would need
not to be freed.
>>"lmfree dataid("did#i")"
>>"lmfree dataid("did#o")"
>>
> Does that free not only both dataids but also both
> allocations?
Yes if the LMINIT wasn't a ddname() one; no if it was a
ddname one.
I mostly used the ddname approach & if my code had done
the allocations it'd then free those (unless what came
next in my code needed the allocations to continue).
--
Jeremy Nicoll - my opinions are my own.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
