On 27 Feb 2013 13:22, Jeff Crosby wrote:
On Wed, Feb 27, 2013 at 4:11 PM, CRPence wrote:
One of the frustrating things about job termination is that the
spooling of the joblog occurs *after* the overrides have been
removed; or, the joblog is produced in a manner that the overrides
in effect for the job do not influence the QPJOBLOG that gets
produced for End-Of-Job (EOJ).
Would not have guessed that, but it makes sense.<<SNIP>>
I accept the effect, though I have never particularly liked it. I
think I understand why it occurs, but I believe it makes little sense,
presuming my understanding is correct.
IIRC the program QMHJLOG runs in the stack of the terminating job,
after a bunch of job-cleanup runs. I am unaware of any reason the
overrides could not remain until after the spooling of the joblog; other
than the likely "that's the way it has always worked, since the
System/38". While all of the opens in the DMCQ would have been closed
first, the DMCQ itself apparently remains, to enable the Common Data
Management open of the QPJOBLOG printer file.? I had long wondered if
changing the destruction of the override information until after the
spooling of QPJOBLOG would be all that the OS would need to do to enable
an override to either QPJOBLOG or *PRTF to remain in effect during that
QMHJLOG processing.? However I understand that such things always seem
easy... until becoming the responsible party for making the changes ;-)
FWiW, Very easy to test\confirm the effect:
sbmjob cmd(ovrprtf qpjoblog ovrscope(*job) hold(*Yes) outq(qezdebug)
copies(3) usrdta(dltme))
log(4 0 *msg) splfacn(*keep) job(myjunkjob) user(*current)
-> Job *n/*n/MYJUNKJOB submitted to job queue .. in library ..
--> the next action is optional, or should by itself be proof
dltsplf *select SELECT(*CURRENT *ALL *ALL dltme *ALL)
-> 0 files deleted. 0 files not deleted.
--> that means I had no spool files with USRDTA(DLTME)
wrkjob myjunkjob option(*SPLF)
-> output from that request:
File Queue User Data Status Pages Page Copies
QPJOBLOG QEZJOBLOG MYJUNKJOB RDY 1 - 1
---> showing none of the overridden attributes in effect :-(
