Our site is going through the project to establish Endevor. One thing they intend to do is load up the Production stage with the source assigned to the proper locations. However there is no intent to generate any thing. No compiling, etc. But the question that was raised was, is there an alternate method to capture the information of the components? We could compile everything if the mainframe had the 'horse power'. But it doesn't. However, one of the thoughts is to generate everything and then remove the output members. Another words capture the compilation info, but then delete the output members (loads). Concern would be that the ACM would have issues with missing members.
Could using CONRELE be an alternate method to preload program / copybook relations .. which mean no compiles.
Send me a note at email@example.com and I'll send you a paper I wrote on something I called the "XPROCESS". It basically gives instruction on doing exactly what you are considering.
There is indeed a method but it involves a little work - so often the easiest approach is to do the compile everything, but NOT replace the final loadlib. Perhaps the easiest way to do this is by modifying your generate processors(s) to test for a specific CCID or COMMENT (e.g. "INITACM") and in that case write the linkedit output to a temporary file and then re-copy the current (GOOD) loadmodule - this will build the ACM info with one extra input component (the copied, original load) that way you can identify component lists that wer built using this migration instead of a normal promotion. Additionally because you are not replacing the produciton load, you can let this process run when ever you have slack time until everything has been processed.
The alternative method is based on the ACM rebuild process. The vanilla version prints all the component list and then rebuilds the ACM info by reading/parsing the components lists - the catch is that you don't have those component lists, but it is possible to 'create' these by a semi-automated process. In principle you will 'scan' your source to find the copy book relationships, and build a 'report' that LOOKS like the component report which you can then feed into the ACM rebuild job.
As I said there is a little work involved to get it right but you can in effect build a complete ACM reference at a fraction of the processing overhead.
Let me know if you need, details, and.or sample processors.
Yes I would like Let me know if you need, details, and.or sample processors.
Samples to address these statements...
"....but it is possible to 'create' these by a semi-automated process. In principle you will 'scan' your source to find the copy book relationships, and build a 'report' that LOOKS like the component report which you can then feed into the ACM rebuild job. ..".
ACM rebuild job?
I will check the manuals for that one.
But if you have a sample, great … or tell me where to find a samples.
This is something we have been doing in CA Services since the Endevor 3.8 release. It involves a scan of Endevor inventory to build ACM information, instead of using Generate actions. It is very low risk, runs fast, and requires no changes to a processor. In fact it does not use a processor.
In one case, it was estimated that it would take weeks to Generate all the programs at a site to build ACM information. This process was tailored for the site and executed in less than 2 days.
1. Upload the .XMIT file as binary (FB80)
2. Receive the .XMIT file to create a PDS
3. Follow the instructions within the Notes document
If anyone prefers to work with a text file, see the .txt.zip attachment in an IEBUPDTE format, and ignore the .XMIT file.
The attachments were updated on 2019 04 23 to: