[fD-bucW] works, but then there is no way to combine the orthogonal and diagonal moves, as there is no shorthand for AD like there is K for WF. So all in all it will not be much shorter.
Interesting is that [fhD-bucK] now does what you want by mistake. Because the K in the second leg forces the XBetza generated by the preprocessor to be based on K, and it currently doesn't make any attempt to transform the fh specification. Which, incidentally, is meaningless on D.
[Edit] I suppose that since it is meaningless, there is no harm in treating fh on a 4-fold atom (and likewise bh, rh, lh) as if we are dealing with an 8-fold atom of the same stride.
[fD-bucW] works, but then there is no way to combine the orthogonal and diagonal moves, as there is no shorthand for AD like there is K for WF. So all in all it will not be much shorter.
Interesting is that [fhD-bucK] now does what you want by mistake. Because the K in the second leg forces the XBetza generated by the preprocessor to be based on K, and it currently doesn't make any attempt to transform the fh specification. Which, incidentally, is meaningless on D.
[Edit] I suppose that since it is meaningless, there is no harm in treating fh on a 4-fold atom (and likewise bh, rh, lh) as if we are dealing with an 8-fold atom of the same stride.