the difference between config and code is largely illusionary, so PHB is right: both can break things and need to be tested.
I would argue that in this case where it was designed like that from the beginning IE detailed as how the system works then there should be no issues.
I did forget to mention it did go though a round of testing, which took less then 5 minutes to be verified as correct. It literally was 6 days of process wrapped around 10 minutes work.
Why stay?
PHB decided that it still counted as a "code" change because someone needed to log onto the box and change the config. Literally it was removing 3 characters from it. The only reason I even had it in the config was to avoid it being a code change to begin with. I argued back that in the time I would take to get through the whole process I could write a screen to do the same thing.
PHB said that would be fine as from that point on its a config not a code change. So off through the whole process, where I present the change to people who don't care, or understand the change and ask for them to approve it. I get that they need awareness, but seriously couldn't a short email saying X is going to happen at time Y suffice?
Im tempted to embed a Python/Ruby interpreter into my next project with a web interface to modify the code. Then I can do these minor changes without going through the 7 day process.