I agree with you that this scenario is unlikely to be the last.
I even think that this might be more often the case than we presume at the moment.
As the Beman project focus on libraries supporting proposals for the C+±standard, it is very likely that there are already matured libraries available which shows how this should work.
I am sure that some maintainers of those libraries kindly agree (as some have already done) that we use their work to start from.
On one hand we could fork from those libraries to make the changes needed for keeping them in sync with the proposal, but this would cause additional effort for 1) the developers (Beman and original) and 2) the library user when they want to change to the standard implementation after successful standardization.
On the other hand we could break compatibility, which would case a lot of hassle for many users which a useful and matured libraries normally have. I am sure that most of the maintainers would strongly dislike this scenario.
Therefore I think that taking some time to consider ways to support 2-interface libraries is needed and well spent.