Whilst I would agree that a bar magnet can be magnetised along its length or across its width, I don't think it is possible to magnetise it in other directions. This is because the magnetising force, if oblique, can be resolved into vectors along and across the magnet. The vector along the magnet will create more flux than that across it, so the direction of magnetisation will be slewed towards the major axis. On the other hand, magnetising the bar across its width is possible because the magnetising force will be at right angles to the long dimension and not have a longitudinal vector. In summary, the applied magnetising force will always be resolved along the axes of symmetry and this will favour the major axis.