[PATCH v2] d3dx9: Simplify math in D3DXSphereBoundProbe

[Alex Henrie]

Signed-off-by: Alex Henrie <alexhenrie24(a)gmail.com>
---
v2: Calculating b isn't actually necessary at all.
---
 dlls/d3dx9_36/mesh.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/dlls/d3dx9_36/mesh.c b/dlls/d3dx9_36/mesh.c
index 5e0bb98efc..381d19a2ee 100644
--- a/dlls/d3dx9_36/mesh.c
+++ b/dlls/d3dx9_36/mesh.c
@@ -2413,16 +2413,13 @@ BOOL WINAPI D3DXSphereBoundProbe(const D3DXVECTOR3 *pcenter, float radius,
         const D3DXVECTOR3 *prayposition, const D3DXVECTOR3 *praydirection)
 {
     D3DXVECTOR3 difference;
-    FLOAT a, b, c, d;
+    FLOAT a, c;
 
     a = D3DXVec3LengthSq(praydirection);
     if (!D3DXVec3Subtract(&difference, prayposition, pcenter)) return FALSE;
-    b = D3DXVec3Dot(&difference, praydirection);
     c = D3DXVec3LengthSq(&difference) - radius * radius;
-    d = b * b - a * c;
 
-    if ( ( d <= 0.0f ) || ( sqrt(d) <= b ) ) return FALSE;
-    return TRUE;
+    return ( a * c < 0 );
 }
 
 /*************************************************************************
-- 
2.27.0

[A]

On 08/07/2020 10:31, Alex Henrie wrote:
> Signed-off-by: Alex Henrie <alexhenrie24(a)gmail.com>
> ---
> v2: Calculating b isn't actually necessary at all.
> ---
>   dlls/d3dx9_36/mesh.c | 7 ++-----
>   1 file changed, 2 insertions(+), 5 deletions(-)
>
> diff --git a/dlls/d3dx9_36/mesh.c b/dlls/d3dx9_36/mesh.c
> index 5e0bb98efc..381d19a2ee 100644
> --- a/dlls/d3dx9_36/mesh.c
> +++ b/dlls/d3dx9_36/mesh.c
> @@ -2413,16 +2413,13 @@ BOOL WINAPI D3DXSphereBoundProbe(const D3DXVECTOR3 *pcenter, float radius,
>           const D3DXVECTOR3 *prayposition, const D3DXVECTOR3 *praydirection)
>   {
>       D3DXVECTOR3 difference;
> -    FLOAT a, b, c, d;
> +    FLOAT a, c;
>
> a = D3DXVec3LengthSq(praydirection);
>       if (!D3DXVec3Subtract(&difference, prayposition, pcenter)) return FALSE;
> -    b = D3DXVec3Dot(&difference, praydirection);
>       c = D3DXVec3LengthSq(&difference) - radius * radius;
> -    d = b * b - a * c;
>
> -    if ( ( d <= 0.0f ) || ( sqrt(d) <= b ) ) return FALSE;
> -    return TRUE;
> +    return ( a * c < 0 );
>   }
>
> /*************************************************************************

It doesn't make much sense to me. You're ignoring the direction of the 
ray. From what I could tell, it seems like you rearranged the right hand 
side of the "or" operator in the existing code, while ignoring the left 
hand side.

Regards,

Ariel

[Matteo Bruni]

On Thu, Jul 9, 2020 at 11:17 AM A <abdaandroid(a)gmail.com> wrote:
> On 08/07/2020 10:31, Alex Henrie wrote:
> > Signed-off-by: Alex Henrie <alexhenrie24(a)gmail.com>
> > ---
> > v2: Calculating b isn't actually necessary at all.
> > ---
> >   dlls/d3dx9_36/mesh.c | 7 ++-----
> >   1 file changed, 2 insertions(+), 5 deletions(-)
> >
> > diff --git a/dlls/d3dx9_36/mesh.c b/dlls/d3dx9_36/mesh.c
> > index 5e0bb98efc..381d19a2ee 100644
> > --- a/dlls/d3dx9_36/mesh.c
> > +++ b/dlls/d3dx9_36/mesh.c
> > @@ -2413,16 +2413,13 @@ BOOL WINAPI D3DXSphereBoundProbe(const D3DXVECTOR3 *pcenter, float radius,
> >           const D3DXVECTOR3 *prayposition, const D3DXVECTOR3 *praydirection)
> >   {
> >       D3DXVECTOR3 difference;
> > -    FLOAT a, b, c, d;
> > +    FLOAT a, c;
> >
> > a = D3DXVec3LengthSq(praydirection);
> >       if (!D3DXVec3Subtract(&difference, prayposition, pcenter)) return FALSE;
> > -    b = D3DXVec3Dot(&difference, praydirection);
> >       c = D3DXVec3LengthSq(&difference) - radius * radius;
> > -    d = b * b - a * c;
> >
> > -    if ( ( d <= 0.0f ) || ( sqrt(d) <= b ) ) return FALSE;
> > -    return TRUE;
> > +    return ( a * c < 0 );
> >   }
> >
> > /*************************************************************************
>
> It doesn't make much sense to me. You're ignoring the direction of the
> ray. From what I could tell, it seems like you rearranged the right hand
> side of the "or" operator in the existing code, while ignoring the left
> hand side.
>
> Regards,
>
> Ariel

True. I'm not sure that the original code made much sense either (e.g.
not normalizing the ray direction vector seems suspicious)... I'm
going to have a proper look at this.