A BBr3 demethylation has been suggested to me as, apparently, it is less prone to attacking double bonds, I am unsure precisely why that may be at present. I am willing to try the BBr3 method, but I don't have any around and I'd have to make some up, which will be a major pain...