0 Replies Latest reply on Oct 8, 2009 2:10 PM by The Flex Viking

    Process Version Control

    The Flex Viking Level 1

      What are most people doing to manage their process version numbers in both a DEV & PROD environment?


      Because an LCA import follows the rule "A process with the same name and version is considered a duplicate and is not imported", it is of course important to make sure any changes to a process in a DEV environment are made on a version number that is different from PROD.


      That suggests 3 main methods.


      1. Easiest in tracking terms is that every process change starts with creating a new process revision.  If this is true, I don't care what the process version is on PROD, the latest process on DEV will always be higher.  The side effect of this is that I could possibly create a TON of unnecessary revisions on DEV and all the headaches that would go along with removing old revisions and whether any of those revisions are RUNNING.
      2. Whenever an LCA is imported to PROD, I go back to DEV and create revisions of all the processes that were just imported.  This ensures that any changes are made against revision which can be imported and I don't have to constantly check to see if a revision is required..  This would take some discipline to enact.
      3. I keep a version table in an XLS that keeps track of what is the top revision on both DEV and PROD.  If I want to update a process, I go check the table to see if I need to create a new revision.  This would require checking the table on every process change.


      My other thought was that I could export the process XML for both the DEV and PROD versions and do a diff.  I was hoping there would be no differences, but in fact, I had a lot of processes whose only difference was the ID on some of the variables.  I am not totally sure why a process imported from an LCA would have different IDs on DEV and PROD, but I can guess it has to do with how LC indentifies all this information in the database.


      So, my thinking is that if the only difference is the ID of some of the variables, the process are in fact identical.


      Any thoughts?