Split the viewer at the y coordinate (which is relative to the
viewer’s surface and must be inside it somewhere) and return the
remaining height. The upper part of the viewer remains (and gets
redrawn on a new surface) and the lower space is now available
for e.g. a new viewer.