Re: d3dx8: patch1: Implement D3DXBoxBoundProbe

[H. Verbeet]

2008/7/27 David Adam :
> -
> ortho = D3DXVec3Dot(praydirection, &normal[i]);
> -
> if ( ortho )
> -
> {
> -
>  t = D3DXVec3Dot(praydirection, &normal[i]) / ortho;

This doesn't make much sense, t will always be close to zero this way.
> - rayposition.x = -3.0f; rayposition.y = 1.0f; rayposition.z = -2.0f;
> - result = D3DXBoxBoundProbe(&bottom_point, &top_point, &rayposition, &raydirection);
> - ok(result == TRUE, "expected TRUE, received FALSE\n");

This succeeds because rayposition lies on one of the faces of the box.
I think it would be useful to also add a test where rayposition lies
outside the box but where the ray still intersects the box.

In general, also note that the fact that the box is always
axis-aligned means the normal is always 0 in two of the three
directions. This simplifies the equations a bit.