On Mon, 14 Sep 2015 09:41:54 -0500, Tom Marchant wrote:
>
>When you code your JCLIN, one thing to remember is that it is coded as
>though the target element is created from the DLIB libraries. In most cases,
>SMP/E doesn't actually build the elements that way, but that's how you tell
>it what needs to be done. Once you understand that, the JCLIN is much
>easier to grasp.
>
But, sometimes extremely frustratingly, not every INCLUDE statement
appearing in Binder JCLIN is issued/executed to build the load module.
This can result in incomplete load modules and the need for Byzantine
circumventions.
>Also, the JCLIN can be simplified from actual JCL that would be required to
>link the load modules or program objects. The data set names provided
>don't have to actually exist. SMP/E uses only the low-level qualifier to
>determine the DDNAME (and DDDEF) to use.
>
//STEP1 EXEC PGM=IEWL
//SYSPRINT DD DSN=WHATEVER.LINKPRT
and:
//STEP2 EXEC PGM=ASMA90
//SYSPRINT DD DSN=WHATEVER.ASMPRT
and have SMP/E examine its DDDEFs and allocate LINKPRT
as SYSPRINT for Binder steps and ASMPRT as SYSPRINT for
assembler steps.
>CALLLIBS processing can be confusing at first, too. Beware of SYSLIB. Both
>the assembler and the binder use SYSLIB, and they use it for very different
>purposes. If your product includes a SYSLIB DDDEF, it will be used for macro
>libraries, not for Auto-call libraries for the binder.
>
Never used CALLLIBS. I need to RTFM. I believe autocall abdicates control
of load module content. Will any SMP/E LIST option show specific versions
(RMIDs) of objects included from CALLLIBS?
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
