Thank you, it's all good and I'm glad it was only a misinterpretation on my part. For what it's worth, I do appreciate the *massive* amount of user engagement choose to do. Thank you for that and for everything else you do for Emacs.
That's a very interesting conclusion, and it makes sense.
Would it make sense to extend the face doc strings that should not use inheritance to indicate that? That combined with the backwards compatibility fix you suggested may very well be a good compromise for the current circumstance.
Alternatively, if there were a path to deprecating header-line and mode-line then the compatibility fix could be skipped.
Thanks,
Aaron