According to the (section 6.1.3) spec, a GetLight() call with a value of position returns the light’s position in eye space coordinates. Is this true for directional lights too? Shouldn’t it return the light’s direction in eye space? I would think that I would want to get (x,y,z) of the light position transformed by the top-left 3x3 inverse-transpose model-view matrix. Instead, according to the spec, it seem that the would get (x,y,z,w=0) of the light position transformed by the 4x4 model-view matrix.
Just curious…