Would anyone who has gone through the process of transitioning their PPM repository to use Git (from any other repository - mine currently uses RCS) be willing to post documentation related to the actual cutover process, tips, pitfalls to avoid, successes, etc.?
I was not working here when it was originally set up to use RCS, however, that being said, I can give what I know of the linkage.
RCS is a standalone repository which we use by updating all Object Types with sufficient commands that when execution of the package line occurs, those commands execute which will check out the appropriate version of the file, stage it where we need it and then either execute it (as a script) or compile it (as in Oracle objects).
Finally, after deployment is completed, custom tables within PPM are updated to indicate the level to which that version of code was just pushed (Dev, Test, Prod, etc.)
At checkin or checkout, we have a form set up to enter the appropriate info (Application, Filename, Version, binary or Ascii, who, is the file locked?, etc) This data is plugged into a couple of tables, KVER_FILES (main with filename and origination info) and KVER_REVISIONS, which relates back to KVER_FILES in a many-to-one manner to allow PPM to have its own copy of the data in the repository. - This keeps these things in synch.
There are various validations in play as well which are application specific like sub-paths etc.
Make use of special commands as well for routine sets of commands that you can then call from the object type command lines. This will keep you from having to duplicate code.
I have included a pic of one of our object type setups so you can at least see the steps of commands involved in an object type. This particular one is for an Oracle Form. I will see if I can pull together anything more for a future dump, but this is a first run-through.