Do the tests really cover everything that should be covered? It seems the negative tests are missing.
What happens if you try to deploy something that should fall (i.e. non step ladder furniture)? Will it refuse to be deployed, and, if not, will it fall as it should?

Also, I assume the chair, which probably was pushed off the ledge (unless the PC was flying) was confirmed to fall as it should (no reason to assume it wouldn't still fall when a message is removed, but still).
Similarly, I assume the step ladder didn't fall.

I'm afraid you're misunderstanding the gist of this PR. This PR is NOT about furniture falling down or not. Non-deployable furniture not falling down was addressed years ago in #47751. Deployable furniture not falling down was addressed in #77083. So, in short, furniture is falling down as it should, both deployable and non-deployable.
This PR is trying to address an issue of debug message showing up when dragging furniture off the ledge.

I did understand what the PR addresses, but I still think it's reasonable to test that there aren't any unexpected side effects, i.e. that the previous behavior remains unchanged.

My tests showed that furniture is falling as it was previously. I didn't notice any unexpected side effects or changes in furniture falling behavior. The only thing that changed is debug message no longer showing up when furniture is dragged off the ledge.