I've thought about it too, but given how many hoops (at least historically) you need to jump through, I'm not sure I want to waste my time. Even if it is on principle.
Also on a note about bug fixing and technical debt, I used to tutor assembly language and I would always recommend people to make sure what they had was error free before moving to the next part of an assignment. As others have said, it's much easier to fix things that are part of your immediate thought process rather than try and figure out what your or someone else did months or years ago.
Working with Unity is a good example of this. I heard from some people I was working with that their networking protocols didn't work at release and they've had to play catch up ever since. If you've tried dealing with inputs in the engine, you'll know how god awful it is. Me and another programmer had to code a dictionary and repeatedly search a massive set of enums thanks to Unity's inability to get the name of the raw input outside of a partially depreciated function.