Re: Create/Last Used Program Dates

[CRPence <CRPbottle@xxxxxxxxx>]

  Every "object" has a creation date in the LIC object header, as well 
every object has a permanent "address" [a "segment"].  When a program is 
"re-compiled" from source as a bound program, the old program object is 
typically renamed and moved to the QRPLOBJ library leaving the original 
object and address with everything except the change date [both rename 
and move update that] and the context [*LIB] in which the object is 
considered to be residing effectively unchanged; i.e. the renamed object 
still has its original creation date.  The new "created" object has its 
own, new, creation date\timestamp.  The "replace object" library named 
QRPLOBJ is cleared on every IPL however, so the older program object was 
presumably long since deleted and thus the address invalidated; since 
the creation date is part of the object and not the pointer, even having 
available a stored copy of the associated pointer can not assist, since 
references to a pointer will effect nothing but MCH3402.  If the OS 
replace-object technique was not used, then that means the deletion was 
just done earlier or perhaps the person\feature doing the "recompile" 
activity had done so; for the former, the same issue as with QRPLOBJ, 
and for the latter, if that moved or renamed copy still exists or was 
saved later, the creation date may still be available from the existing 
object or media.  Because...

  Restore provides a nuanced "creation" in that the LIC header was 
included on the media, and that information replaces\sets the object 
creation date of the new object within the new [segment's] associated 
address.  Thus if there is a backup copy of the *PGM noted in the 
original post for this discussion, if there exists SAVOBJ or SAVLIB 
media that was taken before the last recompile, then the creation 
date\timestamp is still available on that media; i.e. the *PGM could be 
restored into QTEMP, and a DSPOBJD would show the creation information.

Regards, Chuck

On 09-Dec-2011 10:18 , Jerry C. Adams wrote:
> I'm not sure about the object restore, Paul, as I haven't done any
> empirical testing. But I suspect you are correct because CRTDUPOBJ
> and CPYLIB do have a create date = the date of the copy, not the date
> in the original object.
>
> paul therrien  on Friday, December 09, 2011 11:22 AM wrote:
>
> > I may be missing a nuanced conversation here but ...AFAIK the
> > create date is the create date is the create date. The change date
> > may be affected by a couple of operations, but the create date is
> > the compile date. I may be wrong, but I have no reason to believe
> > otherwise.
> >
> > Now, if the program object was placed on the system by a restore
> > command then this may be a different kettle of fish.