Yes, well I forgot to thank you for that pointer! The fixes are in our new
version of 8th. I did notice that the "fp_or" etc words, at least as I saw them, did not perform the binary op on the 'high bytes', e.g. the "excess" ones. So I fixed that in my version of them (since I use tfm, I had to port them anyway).