compilation fixes, disable ubyte color code for now at least

12 files changed, 279 insertions, 493 deletions

diff --git a/src/mesa/drivers/dri/r200/r200_context.c b/src/mesa/drivers/dri/r200/r200_context.c
index ec95f05e6ef..ca6ce4974ae 100644
--- a/src/mesa/drivers/dri/r200/r200_context.c
+++ b/src/mesa/drivers/dri/r200/r200_context.c
@@ -151,10 +151,10 @@ static const char * const card_extensions[] =
     NULL
 };
 
-extern const struct gl_pipeline_stage _r200_render_stage;
-extern const struct gl_pipeline_stage _r200_tcl_stage;
+extern const struct tnl_pipeline_stage _r200_render_stage;
+extern const struct tnl_pipeline_stage _r200_tcl_stage;
 
-static const struct gl_pipeline_stage *r200_pipeline[] = {
+static const struct tnl_pipeline_stage *r200_pipeline[] = {
 
    /* Try and go straight to t&l
     */
diff --git a/src/mesa/drivers/dri/r200/r200_maos_arrays.c b/src/mesa/drivers/dri/r200/r200_maos_arrays.c
index f2abaff3afb..0263d009026 100644
--- a/src/mesa/drivers/dri/r200/r200_maos_arrays.c
+++ b/src/mesa/drivers/dri/r200/r200_maos_arrays.c
@@ -43,7 +43,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "math/m_translate.h"
 #include "tnl/tnl.h"
 #include "tnl/t_context.h"
-#include "tnl/t_imm_debug.h"
 
 #include "r200_context.h"
 #include "r200_ioctl.h"
@@ -316,9 +315,6 @@ void r200EmitArrays( GLcontext *ctx, GLuint inputs )
    GLuint vfmt0 = 0, vfmt1 = 0;
    GLuint count = VB->Count;
    
-   if (R200_DEBUG & DEBUG_VERTS) 
-      _tnl_print_vert_flags( __FUNCTION__, inputs );
-
    if (1) {
       if (!rmesa->tcl.obj.buf) 
 	 emit_vector( ctx, 
@@ -333,7 +329,7 @@ void r200EmitArrays( GLcontext *ctx, GLuint inputs )
       case 3: vfmt0 |= R200_VTX_Z0;
       case 2: 
       default:
-	break;
+	 break;
       }
       component[nr++] = &rmesa->tcl.obj;
    }
@@ -353,60 +349,44 @@ void r200EmitArrays( GLcontext *ctx, GLuint inputs )
    }
 
    if (inputs & VERT_BIT_COLOR0) {
-      if (VB->ColorPtr[0]->Type == GL_UNSIGNED_BYTE) {
-	 if (!rmesa->tcl.rgba.buf)
-	    emit_ubyte_rgba( ctx, 
-			     &rmesa->tcl.rgba, 
-			     (char *)VB->ColorPtr[0]->Ptr,
-			     VB->ColorPtr[0]->Size,
-			     VB->ColorPtr[0]->StrideB,
-			     count);
-
-	 vfmt0 |= R200_VTX_PK_RGBA << R200_VTX_COLOR_0_SHIFT; 
+      int emitsize;
+
+      if (VB->ColorPtr[0]->size == 4 &&
+	  (VB->ColorPtr[0]->stride != 0 ||
+	   VB->ColorPtr[0]->data[0][3] != 1.0)) { 
+	 vfmt0 |= R200_VTX_FP_RGBA << R200_VTX_COLOR_0_SHIFT; 
+	 emitsize = 4;
       }
-      else {
-	 int emitsize;
-
-	 if (VB->ColorPtr[0]->Size == 4 &&
-	     (VB->ColorPtr[0]->StrideB != 0 ||
-	      ((GLfloat *)VB->ColorPtr[0]->Ptr)[3] != 1.0)) { 
-	    vfmt0 |= R200_VTX_FP_RGBA << R200_VTX_COLOR_0_SHIFT; 
-	    emitsize = 4;
-	 }
-	 else { 
-	    vfmt0 |= R200_VTX_FP_RGB << R200_VTX_COLOR_0_SHIFT; 
-	    emitsize = 3;
-	 }
-
-	 if (!rmesa->tcl.rgba.buf)
-	    emit_vector( ctx, 
-			 &(rmesa->tcl.rgba), 
-			 (char *)VB->ColorPtr[0]->Ptr,
-			 emitsize,
-			 VB->ColorPtr[0]->StrideB,
-			 count);
+      else { 
+	 vfmt0 |= R200_VTX_FP_RGB << R200_VTX_COLOR_0_SHIFT; 
+	 emitsize = 3;
       }
 
+      if (!rmesa->tcl.rgba.buf)
+	 emit_vector( ctx, 
+		      &(rmesa->tcl.rgba), 
+		      (char *)VB->ColorPtr[0]->data,
+		      emitsize,
+		      VB->ColorPtr[0]->stride,
+		      count);
+
       component[nr++] = &rmesa->tcl.rgba;
    }
 
 
    if (inputs & VERT_BIT_COLOR1) {
       if (!rmesa->tcl.spec.buf) {
-	 if (VB->SecondaryColorPtr[0]->Type != GL_UNSIGNED_BYTE)
-	    r200_import_float_spec_colors( ctx );
-
-	 emit_ubyte_rgba( ctx, 
-			  &rmesa->tcl.spec, 
-			  (char *)VB->SecondaryColorPtr[0]->Ptr,
-			  3,
-			  VB->SecondaryColorPtr[0]->StrideB,
-			  count);
+	 emit_vector( ctx, 
+		      &rmesa->tcl.spec, 
+		      (char *)VB->SecondaryColorPtr[0]->data,
+		      3,
+		      VB->SecondaryColorPtr[0]->stride,
+		      count);
       }
 
       /* How does this work?
        */
-      vfmt0 |= R200_VTX_PK_RGBA << R200_VTX_COLOR_1_SHIFT; 
+      vfmt0 |= R200_VTX_FP_RGB << R200_VTX_COLOR_1_SHIFT; 
       component[nr++] = &rmesa->tcl.spec;
    }
 
@@ -455,8 +435,8 @@ void r200ReleaseArrays( GLcontext *ctx, GLuint newinputs )
 {
    r200ContextPtr rmesa = R200_CONTEXT( ctx );
 
-   if (R200_DEBUG & DEBUG_VERTS) 
-      _tnl_print_vert_flags( __FUNCTION__, newinputs );
+/*    if (R200_DEBUG & DEBUG_VERTS)  */
+/*       _tnl_print_vert_flags( __FUNCTION__, newinputs ); */
 
    if (newinputs & VERT_BIT_POS) 
      r200ReleaseDmaRegion( rmesa, &rmesa->tcl.obj, __FUNCTION__ );
diff --git a/src/mesa/drivers/dri/r200/r200_state.c b/src/mesa/drivers/dri/r200/r200_state.c
index 55a00cd4de7..27dc6fd1f2b 100644
--- a/src/mesa/drivers/dri/r200/r200_state.c
+++ b/src/mesa/drivers/dri/r200/r200_state.c
@@ -2089,10 +2089,30 @@ static void r200InvalidateState( GLcontext *ctx, GLuint new_state )
    r200VtxfmtInvalidate( ctx );
 }
 
+/* A hack.  The r200 can actually cope just fine with materials
+ * between begin/ends, so fix this.
+ */
+static GLboolean check_material( GLcontext *ctx )
+{
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   GLint i;
+
+   for (i = _TNL_ATTRIB_MAT_FRONT_AMBIENT; 
+	i < _TNL_ATTRIB_MAT_BACK_INDEXES; 
+	i++)
+      if (tnl->vb.AttribPtr[i] &&
+	  tnl->vb.AttribPtr[i]->stride)
+	 return GL_TRUE;
+
+   return GL_FALSE;
+}
+      
+
 static void r200WrapRunPipeline( GLcontext *ctx )
 {
    r200ContextPtr rmesa = R200_CONTEXT(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
+   GLboolean has_material;
 
    if (0)
       fprintf(stderr, "%s, newstate: %x\n", __FUNCTION__, rmesa->NewGLState);
@@ -2102,7 +2122,9 @@ static void r200WrapRunPipeline( GLcontext *ctx )
    if (rmesa->NewGLState)
       r200ValidateState( ctx );
 
-   if (tnl->vb.Material) {
+   has_material = (ctx->Light.Enabled && check_material( ctx ));
+
+   if (has_material) {
       TCL_FALLBACK( ctx, R200_TCL_FALLBACK_MATERIAL, GL_TRUE );
    }
 
@@ -2110,7 +2132,7 @@ static void r200WrapRunPipeline( GLcontext *ctx )
     */ 
    _tnl_run_pipeline( ctx );
 
-   if (tnl->vb.Material) {
+   if (has_material) {
       TCL_FALLBACK( ctx, R200_TCL_FALLBACK_MATERIAL, GL_FALSE );
       r200UpdateMaterial( ctx ); /* not needed any more? */
    }
diff --git a/src/mesa/drivers/dri/r200/r200_swtcl.c b/src/mesa/drivers/dri/r200/r200_swtcl.c
index dd4025235b6..48b438cc292 100644
--- a/src/mesa/drivers/dri/r200/r200_swtcl.c
+++ b/src/mesa/drivers/dri/r200/r200_swtcl.c
@@ -48,7 +48,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "math/m_translate.h"
 #include "tnl/tnl.h"
 #include "tnl/t_context.h"
-#include "tnl/t_imm_exec.h"
 #include "tnl/t_pipeline.h"
 
 #include "r200_context.h"
@@ -538,7 +537,7 @@ static void VERT_FALLBACK( GLcontext *ctx,
    tnl->Driver.Render.PrimitiveNotify( ctx, flags & PRIM_MODE_MASK );
    tnl->Driver.Render.BuildVertices( ctx, start, count, ~0 );
    tnl->Driver.Render.PrimTabVerts[flags&PRIM_MODE_MASK]( ctx, start, count, flags );
-   R200_CONTEXT(ctx)->swtcl.SetupNewInputs = VERT_BIT_CLIP;
+   R200_CONTEXT(ctx)->swtcl.SetupNewInputs = _TNL_BIT_POS;
 }
 
 static void ELT_FALLBACK( GLcontext *ctx,
@@ -550,7 +549,7 @@ static void ELT_FALLBACK( GLcontext *ctx,
    tnl->Driver.Render.PrimitiveNotify( ctx, flags & PRIM_MODE_MASK );
    tnl->Driver.Render.BuildVertices( ctx, start, count, ~0 );
    tnl->Driver.Render.PrimTabElts[flags&PRIM_MODE_MASK]( ctx, start, count, flags );
-   R200_CONTEXT(ctx)->swtcl.SetupNewInputs = VERT_BIT_CLIP;
+   R200_CONTEXT(ctx)->swtcl.SetupNewInputs = _TNL_BIT_POS;
 }
 
 
@@ -632,7 +631,7 @@ do {									\
 
 
 static GLboolean r200_run_render( GLcontext *ctx,
-				    struct gl_pipeline_stage *stage )
+				    struct tnl_pipeline_stage *stage )
 {
    r200ContextPtr rmesa = R200_CONTEXT(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
@@ -660,18 +659,22 @@ static GLboolean r200_run_render( GLcontext *ctx,
 
    tnl->Driver.Render.Start( ctx );
 
-   for (i = 0 ; !(flags & PRIM_LAST) ; i += length)
+   for (i = 0 ; i < VB->PrimitiveCount ; i++)
    {
-      flags = VB->Primitive[i];
-      length = VB->PrimitiveLength[i];
+      GLuint prim = VB->Primitive[i].mode;
+      GLuint start = VB->Primitive[i].start;
+      GLuint length = VB->Primitive[i].count;
+
+      if (!length)
+	 continue;
 
       if (R200_DEBUG & DEBUG_PRIMS)
 	 fprintf(stderr, "r200_render.c: prim %s %d..%d\n", 
-		 _mesa_lookup_enum_by_nr(flags & PRIM_MODE_MASK), 
-		 i, i+length);
+		 _mesa_lookup_enum_by_nr(prim & PRIM_MODE_MASK), 
+		 start, start+length);
 
       if (length)
-	 tab[flags & PRIM_MODE_MASK]( ctx, i, i + length, flags );
+	 tab[prim & PRIM_MODE_MASK]( ctx, start, start + length, flags );
    }
 
    tnl->Driver.Render.Finish( ctx );
@@ -682,35 +685,35 @@ static GLboolean r200_run_render( GLcontext *ctx,
 
 
 static void r200_check_render( GLcontext *ctx,
-				 struct gl_pipeline_stage *stage )
+				 struct tnl_pipeline_stage *stage )
 {
-   GLuint inputs = VERT_BIT_POS | VERT_BIT_CLIP | VERT_BIT_COLOR0;
+   GLuint inputs = _TNL_BIT_POS | _TNL_BIT_COLOR0;
 
    if (ctx->RenderMode == GL_RENDER) {
       if (ctx->_TriangleCaps & DD_SEPARATE_SPECULAR)
-	 inputs |= VERT_BIT_COLOR1;
+	 inputs |= _TNL_BIT_COLOR1;
 
       if (ctx->Texture.Unit[0]._ReallyEnabled)
-	 inputs |= VERT_BIT_TEX0;
+	 inputs |= _TNL_BIT_TEX0;
 
       if (ctx->Texture.Unit[1]._ReallyEnabled)
-	 inputs |= VERT_BIT_TEX1;
+	 inputs |= _TNL_BIT_TEX1;
 
       if (ctx->Fog.Enabled)
-	 inputs |= VERT_BIT_FOG;
+	 inputs |= _TNL_BIT_FOG;
    }
 
    stage->inputs = inputs;
 }
 
 
-static void dtr( struct gl_pipeline_stage *stage )
+static void dtr( struct tnl_pipeline_stage *stage )
 {
    (void)stage;
 }
 
 
-const struct gl_pipeline_stage _r200_render_stage =
+const struct tnl_pipeline_stage _r200_render_stage =
 {
    "r200 render",
    (_DD_NEW_SEPARATE_SPECULAR |
@@ -1231,7 +1234,7 @@ r200PointsBitmap( GLcontext *ctx, GLint px, GLint py,
 
 void r200FlushVertices( GLcontext *ctx, GLuint flags )
 {
-   _tnl_flush_vertices( ctx, flags );
+   _tnl_FlushVertices( ctx, flags );
 
    if (flags & FLUSH_STORED_VERTICES)
       R200_NEWPRIM( R200_CONTEXT( ctx ) );
diff --git a/src/mesa/drivers/dri/r200/r200_tcl.c b/src/mesa/drivers/dri/r200/r200_tcl.c
index 9869ad61963..71e3d5256a9 100644
--- a/src/mesa/drivers/dri/r200/r200_tcl.c
+++ b/src/mesa/drivers/dri/r200/r200_tcl.c
@@ -255,7 +255,7 @@ void r200TclPrimitive( GLcontext *ctx,
 /* TCL render.
  */
 static GLboolean r200_run_tcl_render( GLcontext *ctx,
-					struct gl_pipeline_stage *stage )
+				      struct tnl_pipeline_stage *stage )
 {
    r200ContextPtr rmesa = R200_CONTEXT(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
@@ -278,24 +278,19 @@ static GLboolean r200_run_tcl_render( GLcontext *ctx,
 
    rmesa->tcl.Elts = VB->Elts;
 
-   for (i = VB->FirstPrimitive ; !(flags & PRIM_LAST) ; i += length)
+   for (i = 0 ; i < VB->PrimitiveCount ; i++)
    {
-      flags = VB->Primitive[i];
-      length = VB->PrimitiveLength[i];
-
-      if (R200_DEBUG & DEBUG_PRIMS)
-	 fprintf(stderr, "%s: prim %s %d..%d\n", 
-		 __FUNCTION__,
-		 _mesa_lookup_enum_by_nr(flags & PRIM_MODE_MASK), 
-		 i, i+length);
+      GLuint prim = VB->Primitive[i].mode;
+      GLuint start = VB->Primitive[i].start;
+      GLuint length = VB->Primitive[i].count;
 
       if (!length)
 	 continue;
 
       if (rmesa->tcl.Elts)
-	 r200EmitEltPrimitive( ctx, i, i+length, flags );
+	 r200EmitEltPrimitive( ctx, start, start+length, prim );
       else
-	 r200EmitPrimitive( ctx, i, i+length, flags );
+	 r200EmitPrimitive( ctx, start, start+length, prim );
    }
 
    return GL_FALSE;		/* finished the pipe */
@@ -304,7 +299,7 @@ static GLboolean r200_run_tcl_render( GLcontext *ctx,
 
 
 static void r200_check_tcl_render( GLcontext *ctx,
-				     struct gl_pipeline_stage *stage )
+				   struct tnl_pipeline_stage *stage )
 {
    r200ContextPtr rmesa = R200_CONTEXT(ctx);
    GLuint inputs = VERT_BIT_POS;
@@ -360,13 +355,13 @@ static void r200_check_tcl_render( GLcontext *ctx,
 }
 
 static void r200_init_tcl_render( GLcontext *ctx,
-				    struct gl_pipeline_stage *stage )
+				    struct tnl_pipeline_stage *stage )
 {
    stage->check = r200_check_tcl_render;
    stage->check( ctx, stage );
 }
 
-static void dtr( struct gl_pipeline_stage *stage )
+static void dtr( struct tnl_pipeline_stage *stage )
 {
    (void)stage;
 }
@@ -374,7 +369,7 @@ static void dtr( struct gl_pipeline_stage *stage )
 
 /* Initial state for tcl stage.  
  */
-const struct gl_pipeline_stage _r200_tcl_stage =
+const struct tnl_pipeline_stage _r200_tcl_stage =
 {
    "r200 render",
    (_DD_NEW_SEPARATE_SPECULAR |
diff --git a/src/mesa/drivers/dri/r200/r200_vtxfmt.c b/src/mesa/drivers/dri/r200/r200_vtxfmt.c
index 5a4f59d9763..824216eb350 100644
--- a/src/mesa/drivers/dri/r200/r200_vtxfmt.c
+++ b/src/mesa/drivers/dri/r200/r200_vtxfmt.c
@@ -75,10 +75,10 @@ static void count_funcs( r200ContextPtr rmesa )
    count_func( "Vertex2fv", &rmesa->vb.dfn_cache.Vertex2fv );
    count_func( "Vertex3f", &rmesa->vb.dfn_cache.Vertex3f );
    count_func( "Vertex3fv", &rmesa->vb.dfn_cache.Vertex3fv );
-   count_func( "Color4ub", &rmesa->vb.dfn_cache.Color4ub );
-   count_func( "Color4ubv", &rmesa->vb.dfn_cache.Color4ubv );
-   count_func( "Color3ub", &rmesa->vb.dfn_cache.Color3ub );
-   count_func( "Color3ubv", &rmesa->vb.dfn_cache.Color3ubv );
+   count_func( "Color4ub", &rmesa->vb.dfn_cache.Color4ub ); 
+   count_func( "Color4ubv", &rmesa->vb.dfn_cache.Color4ubv ); 
+   count_func( "Color3ub", &rmesa->vb.dfn_cache.Color3ub ); 
+   count_func( "Color3ubv", &rmesa->vb.dfn_cache.Color3ubv ); 
    count_func( "Color4f", &rmesa->vb.dfn_cache.Color4f );
    count_func( "Color4fv", &rmesa->vb.dfn_cache.Color4fv );
    count_func( "Color3f", &rmesa->vb.dfn_cache.Color3f );
@@ -468,7 +468,10 @@ static void VFMT_FALLBACK( const char *caller )
       glNormal3fv( rmesa->vb.normalptr );
 
    if (VTX_COLOR(ind0, 0) == R200_VTX_PK_RGBA) 
-         glColor4ub( rmesa->vb.colorptr->red, rmesa->vb.colorptr->green, rmesa->vb.colorptr->blue, rmesa->vb.colorptr->alpha );
+         glColor4ub( rmesa->vb.colorptr->red,
+		     rmesa->vb.colorptr->green,
+		     rmesa->vb.colorptr->blue,
+		     rmesa->vb.colorptr->alpha );
    else if (VTX_COLOR(ind0, 0) == R200_VTX_FP_RGBA) 
       glColor4fv( rmesa->vb.floatcolorptr );
    else if (VTX_COLOR(ind0, 0) == R200_VTX_FP_RGB) {
@@ -482,7 +485,9 @@ static void VFMT_FALLBACK( const char *caller )
    }
 
    if (VTX_COLOR(ind0, 1) == R200_VTX_PK_RGBA) 
-      _glapi_Dispatch->SecondaryColor3ubEXT( rmesa->vb.specptr->red, rmesa->vb.specptr->green, rmesa->vb.specptr->blue ); 
+      _glapi_Dispatch->SecondaryColor3ubEXT( rmesa->vb.specptr->red, 
+					     rmesa->vb.specptr->green,
+					     rmesa->vb.specptr->blue ); 
 
    if (ind1 & (7 << R200_VTX_TEX0_COMP_CNT_SHIFT)) 
       glTexCoord2fv( rmesa->vb.texcoordptr[0] );
@@ -1006,13 +1011,14 @@ void r200VtxfmtInit( GLcontext *ctx )
    vfmt->FogCoordfEXT = _mesa_noop_FogCoordfEXT;
    vfmt->EdgeFlag = _mesa_noop_EdgeFlag;
    vfmt->EdgeFlagv = _mesa_noop_EdgeFlagv;
-   vfmt->Indexi = _mesa_noop_Indexi;
-   vfmt->Indexiv = _mesa_noop_Indexiv;
+   vfmt->Indexf = _mesa_noop_Indexf;
+   vfmt->Indexfv = _mesa_noop_Indexfv;
 
 
    /* Active but unsupported -- fallback if we receive these:
     */
    vfmt->CallList = r200_fallback_CallList;
+   vfmt->CallLists = r200_fallback_CallLists;
    vfmt->EvalCoord1f = r200_fallback_EvalCoord1f;
    vfmt->EvalCoord1fv = r200_fallback_EvalCoord1fv;
    vfmt->EvalCoord2f = r200_fallback_EvalCoord2f;
diff --git a/src/mesa/drivers/dri/r200/r200_vtxfmt_c.c b/src/mesa/drivers/dri/r200/r200_vtxfmt_c.c
index b7f3fec43e0..bdda60eea09 100644
--- a/src/mesa/drivers/dri/r200/r200_vtxfmt_c.c
+++ b/src/mesa/drivers/dri/r200/r200_vtxfmt_c.c
@@ -827,16 +827,10 @@ void r200VtxfmtInitChoosers( GLvertexformat *vfmt )
 {
    vfmt->Color3f = choose_Color3f;
    vfmt->Color3fv = choose_Color3fv;
-   vfmt->Color3ub = choose_Color3ub;
-   vfmt->Color3ubv = choose_Color3ubv;
    vfmt->Color4f = choose_Color4f;
    vfmt->Color4fv = choose_Color4fv;
-   vfmt->Color4ub = choose_Color4ub;
-   vfmt->Color4ubv = choose_Color4ubv;
    vfmt->SecondaryColor3fEXT = choose_SecondaryColor3fEXT;
    vfmt->SecondaryColor3fvEXT = choose_SecondaryColor3fvEXT;
-   vfmt->SecondaryColor3ubEXT = choose_SecondaryColor3ubEXT;
-   vfmt->SecondaryColor3ubvEXT = choose_SecondaryColor3ubvEXT;
    vfmt->MultiTexCoord1fARB = choose_MultiTexCoord1fARB;
    vfmt->MultiTexCoord1fvARB = choose_MultiTexCoord1fvARB;
    vfmt->MultiTexCoord2fARB = choose_MultiTexCoord2fARB;
@@ -851,6 +845,17 @@ void r200VtxfmtInitChoosers( GLvertexformat *vfmt )
    vfmt->Vertex2fv = choose_Vertex2fv;
    vfmt->Vertex3f = choose_Vertex3f;
    vfmt->Vertex3fv = choose_Vertex3fv;
+
+   /* TODO: restore ubyte colors to vtxfmt.
+    */
+#if 0
+   vfmt->Color3ub = choose_Color3ub;
+   vfmt->Color3ubv = choose_Color3ubv;
+   vfmt->Color4ub = choose_Color4ub;
+   vfmt->Color4ubv = choose_Color4ubv;
+   vfmt->SecondaryColor3ubEXT = choose_SecondaryColor3ubEXT;
+   vfmt->SecondaryColor3ubvEXT = choose_SecondaryColor3ubvEXT;
+#endif
 }
 
 
diff --git a/src/mesa/tnl_dd/t_dd_dmatmp.h b/src/mesa/tnl_dd/t_dd_dmatmp.h
index d9f709389d1..10ff1df5f0b 100644
--- a/src/mesa/tnl_dd/t_dd_dmatmp.h
+++ b/src/mesa/tnl_dd/t_dd_dmatmp.h
@@ -318,14 +318,6 @@ static void TAG(render_tri_strip_verts)( GLcontext *ctx,
 	 currentsz = dmasz;
       }
 
-      if ((flags & PRIM_PARITY) && count - start > 2) {
-	 if (HAVE_TRI_STRIP_1 && 0) {
-	 } else {
-	    EMIT_VERTS( ctx, start, 1 );
-	    currentsz--;
-	 }
-      }
-
       /* From here on emit even numbers of tris when wrapping over buffers:
        */
       dmasz -= (dmasz & 1);
@@ -869,10 +861,6 @@ static void TAG(render_tri_strip_elts)( GLcontext *ctx,
 	 currentsz = dmasz;
       }
 
-      if ((flags & PRIM_PARITY) && count - start > 2) {
-	 TAG(emit_elts)( ctx, elts+start, 1 );
-      }
-
       /* Keep the same winding over multiple buffers:
        */
       dmasz -= (dmasz & 1);
diff --git a/src/mesa/tnl_dd/t_dd_dmatmp2.h b/src/mesa/tnl_dd/t_dd_dmatmp2.h
index 3f58d24c952..dd0286b453b 100644
--- a/src/mesa/tnl_dd/t_dd_dmatmp2.h
+++ b/src/mesa/tnl_dd/t_dd_dmatmp2.h
@@ -323,9 +323,6 @@ static void TAG(render_tri_strip_verts)( GLcontext *ctx,
 
       ELT_INIT( GL_TRIANGLES, HW_TRIANGLES );
 
-      if (flags & PRIM_PARITY)
-	 parity = 1;
-
       /* Emit even number of tris in each full buffer.
        */
       dmasz = dmasz/3;
@@ -348,8 +345,6 @@ static void TAG(render_tri_strip_verts)( GLcontext *ctx,
 	 CLOSE_ELTS();
       }
    }
-   else if ((flags & PRIM_PARITY) == 0)  
-      EMIT_PRIM( ctx, GL_TRIANGLE_STRIP, HW_TRIANGLE_STRIP_0, start, count );
    else if (HAVE_TRI_STRIP_1)
       EMIT_PRIM( ctx, GL_TRIANGLE_STRIP, HW_TRIANGLE_STRIP_1, start, count );
    else {
@@ -757,18 +752,9 @@ static void TAG(render_tri_strip_elts)( GLcontext *ctx,
    for (j = start ; j + 2 < count; j += nr - 2 ) {
       nr = MIN2( dmasz, count - j );
 
-      if (flags & PRIM_PARITY) {
-	 dest = ALLOC_ELTS( nr );	
-	 dest = TAG(emit_elts)( ctx, dest, elts+j, 1 );
-	 dest = TAG(emit_elts)( ctx, dest, elts+j, nr-1 );
-	 nr--; flags &= ~PRIM_PARITY;
-	 CLOSE_ELTS();
-      }
-      else {
-	 dest = ALLOC_ELTS( nr );
-	 dest = TAG(emit_elts)( ctx, dest, elts+j, nr );
-	 CLOSE_ELTS();
-      }
+      dest = ALLOC_ELTS( nr );
+      dest = TAG(emit_elts)( ctx, dest, elts+j, nr );
+      CLOSE_ELTS();
    }
 }
 
diff --git a/src/mesa/tnl_dd/t_dd_tritmp.h b/src/mesa/tnl_dd/t_dd_tritmp.h
index e6e90170e9c..b26347b77eb 100644
--- a/src/mesa/tnl_dd/t_dd_tritmp.h
+++ b/src/mesa/tnl_dd/t_dd_tritmp.h
@@ -176,9 +176,8 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
 		  }
 	       }
 	       else {
-		  GLchan (*vbcolor)[4] = (GLchan (*)[4])VB->ColorPtr[1]->Ptr;
-		  ASSERT(VB->ColorPtr[1]->Type == CHAN_TYPE);
-		  ASSERT(VB->ColorPtr[1]->StrideB == 4*sizeof(GLchan));
+		  GLfloat (*vbcolor)[4] = VB->ColorPtr[1]->data;
+		  ASSERT(VB->ColorPtr[1]->stride == 4*sizeof(GLfloat));
 		  (void) vbcolor;
 
 		  if (!DO_FLAT) {
@@ -191,7 +190,7 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
 		  VERT_SET_RGBA( v[2], vbcolor[e2] );
 
 		  if (HAVE_SPEC && VB->SecondaryColorPtr[1]) {
-		     GLchan (*vbspec)[4] = (GLchan (*)[4])VB->SecondaryColorPtr[1]->Ptr;
+		     GLfloat (*vbspec)[4] = VB->SecondaryColorPtr[1]->data;
 
 		     if (!DO_FLAT) {
 			VERT_SAVE_SPEC( 0 );
@@ -205,11 +204,14 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
 	       }
 	    }
 	    else {
-	       GLfloat *vbindex = VB->IndexPtr[1]->data;
+	       GLfloat (*vbindex) = (GLfloat *)VB->IndexPtr[1]->data;
 	       if (!DO_FLAT) {
+		  VERT_SAVE_IND( 0 );
+		  VERT_SAVE_IND( 1 );
 		  VERT_SET_IND( v[0], vbindex[e0] );
 		  VERT_SET_IND( v[1], vbindex[e1] );
 	       }
+	       VERT_SAVE_IND( 2 );
 	       VERT_SET_IND( v[2], vbindex[e2] );
 	    }
 	 }
@@ -308,12 +310,11 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
 	 }
       }
       else {
-	 GLfloat *vbindex = VB->IndexPtr[0]->data;
 	 if (!DO_FLAT) {
-	    VERT_SET_IND( v[0], vbindex[e0] );
-	    VERT_SET_IND( v[1], vbindex[e1] );
+	    VERT_RESTORE_IND( 0 );
+	    VERT_RESTORE_IND( 1 );
 	 }
-	 VERT_SET_IND( v[2], vbindex[e2] );
+	 VERT_RESTORE_IND( 2 );
       }
    }
 
@@ -384,7 +385,7 @@ static void TAG(quad)( GLcontext *ctx,
 	 if (DO_TWOSIDE && facing == 1)
 	 {
 	    if (HAVE_RGBA) {
-	       GLchan (*vbcolor)[4] = (GLchan (*)[4])VB->ColorPtr[1]->Ptr;
+	       GLfloat (*vbcolor)[4] = VB->ColorPtr[1]->data;
 	       (void)vbcolor;
 
 	       if (HAVE_BACK_COLORS) {
@@ -424,8 +425,8 @@ static void TAG(quad)( GLcontext *ctx,
 	          VERT_SET_RGBA( v[3], vbcolor[e3] );
 
 	          if (HAVE_SPEC && VB->SecondaryColorPtr[1]) {
-		     GLchan (*vbspec)[4] = (GLchan (*)[4])VB->SecondaryColorPtr[1]->Ptr;
-		     ASSERT(VB->SecondaryColorPtr[1]->StrideB==4*sizeof(GLchan));
+		     GLfloat (*vbspec)[4] = VB->SecondaryColorPtr[1]->data;
+		     ASSERT(VB->SecondaryColorPtr[1]->stride==4*sizeof(GLfloat));
 
 		     if (!DO_FLAT) {
 		        VERT_SAVE_SPEC( 0 );
@@ -441,12 +442,16 @@ static void TAG(quad)( GLcontext *ctx,
 	       }
 	    }
 	    else {
-	       GLfloat *vbindex = VB->IndexPtr[1]->data;
+	       GLfloat *vbindex = (GLfloat *)VB->IndexPtr[1]->data;
 	       if (!DO_FLAT) {
+		  VERT_SAVE_IND( 0 );
+		  VERT_SAVE_IND( 1 );
+		  VERT_SAVE_IND( 2 );
 		  VERT_SET_IND( v[0], vbindex[e0] );
 		  VERT_SET_IND( v[1], vbindex[e1] );
 		  VERT_SET_IND( v[2], vbindex[e2] );
 	       }
+	       VERT_SAVE_IND( 3 );
 	       VERT_SET_IND( v[3], vbindex[e3] );
 	    }
 	 }
@@ -557,13 +562,12 @@ static void TAG(quad)( GLcontext *ctx,
 	 }
       }
       else {
-	 GLfloat *vbindex = VB->IndexPtr[0]->data;
 	 if (!DO_FLAT) {
-	    VERT_SET_IND( v[0], vbindex[e0] );
-	    VERT_SET_IND( v[1], vbindex[e1] );
-	    VERT_SET_IND( v[2], vbindex[e2] );
+	    VERT_RESTORE_IND( 0 );
+	    VERT_RESTORE_IND( 1 );
+	    VERT_RESTORE_IND( 2 );
 	 }
-	 VERT_SET_IND( v[3], vbindex[e3] );
+	 VERT_RESTORE_IND( 3 );
       }
    }
 
@@ -611,7 +615,7 @@ static void TAG(quad)( GLcontext *ctx, GLuint e0,
 #if DO_LINE
 static void TAG(line)( GLcontext *ctx, GLuint e0, GLuint e1 )
 {
-   TNLvertexbuffer *VB = &TNL_CONTEXT(ctx)->vb;
+   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
    VERTEX *v[2];
    LOCAL_VARS(2);
 
diff --git a/src/mesa/tnl_dd/t_dd_vb.c b/src/mesa/tnl_dd/t_dd_vb.c
index 4a742bcbbee..380b0424b4f 100644
--- a/src/mesa/tnl_dd/t_dd_vb.c
+++ b/src/mesa/tnl_dd/t_dd_vb.c
@@ -271,56 +271,7 @@ void TAG(print_vertex)( GLcontext *ctx, const VERTEX *v )
    fprintf(stderr, "\n");
 }
 
-static void do_import( struct vertex_buffer *VB,
-		       struct gl_client_array *to,
-		       struct gl_client_array *from )
-{
-   GLuint count = VB->Count;
-
-   if (!to->Ptr) {
-      to->Ptr = ALIGN_MALLOC( VB->Size * 4 * sizeof(GLubyte), 32 );
-      to->Type = GL_UNSIGNED_BYTE;
-   }
-
-   /* No need to transform the same value 3000 times.
-    */
-   if (!from->StrideB) {
-      to->StrideB = 0;
-      count = 1;
-   }
-   else
-      to->StrideB = 4 * sizeof(GLubyte);
-   
-   _math_trans_4ub( (GLubyte (*)[4]) to->Ptr,
-		    from->Ptr,
-		    from->StrideB,
-		    from->Type,
-		    from->Size,
-		    0,
-		    count);
-}
-
-#ifndef IMPORT_QUALIFIER
-#define IMPORT_QUALIFIER static
-#endif
-
-IMPORT_QUALIFIER void TAG(import_float_colors)( GLcontext *ctx )
-{
-   LOCALVARS
-   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
-   struct gl_client_array *to = GET_UBYTE_COLOR_STORE();
-   do_import( VB, to, VB->ColorPtr[0] );
-   VB->ColorPtr[0] = to;
-}
 
-IMPORT_QUALIFIER void TAG(import_float_spec_colors)( GLcontext *ctx )
-{
-   LOCALVARS
-   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
-   struct gl_client_array *to = GET_UBYTE_SPEC_COLOR_STORE();
-   do_import( VB, to, VB->SecondaryColorPtr[0] );
-   VB->SecondaryColorPtr[0] = to;
-}
 
 /* Interpolate the elements of the VB not included in typical hardware
  * vertices.  
@@ -332,7 +283,7 @@ IMPORT_QUALIFIER void TAG(import_float_spec_colors)( GLcontext *ctx )
 #define INTERP_QUALIFIER static
 #endif
 
-#define GET_COLOR(ptr, idx) (((GLchan (*)[4])((ptr)->Ptr))[idx])
+#define GET_COLOR(ptr, idx) ((ptr)->data[idx])
 
 
 INTERP_QUALIFIER void TAG(interp_extras)( GLcontext *ctx,
@@ -344,13 +295,15 @@ INTERP_QUALIFIER void TAG(interp_extras)( GLcontext *ctx,
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
    if (VB->ColorPtr[1]) {
-      INTERP_4CHAN( t,
+      assert(VB->ColorPtr[1]->stride == 4 * sizeof(GLfloat));
+
+      INTERP_4F( t,
 		    GET_COLOR(VB->ColorPtr[1], dst),
 		    GET_COLOR(VB->ColorPtr[1], out),
 		    GET_COLOR(VB->ColorPtr[1], in) );
 
       if (VB->SecondaryColorPtr[1]) {
-	 INTERP_3CHAN( t,
+	 INTERP_3F( t,
 		       GET_COLOR(VB->SecondaryColorPtr[1], dst),
 		       GET_COLOR(VB->SecondaryColorPtr[1], out),
 		       GET_COLOR(VB->SecondaryColorPtr[1], in) );
@@ -371,12 +324,12 @@ INTERP_QUALIFIER void TAG(copy_pv_extras)( GLcontext *ctx,
       struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
    if (VB->ColorPtr[1]) {
-      COPY_CHAN4( GET_COLOR(VB->ColorPtr[1], dst), 
-		  GET_COLOR(VB->ColorPtr[1], src) );
+      COPY_4FV( GET_COLOR(VB->ColorPtr[1], dst), 
+		GET_COLOR(VB->ColorPtr[1], src) );
 
       if (VB->SecondaryColorPtr[1]) {
-	 COPY_CHAN4( GET_COLOR(VB->SecondaryColorPtr[1], dst), 
-		     GET_COLOR(VB->SecondaryColorPtr[1], src) );
+	 COPY_4FV( GET_COLOR(VB->SecondaryColorPtr[1], dst), 
+		   GET_COLOR(VB->SecondaryColorPtr[1], src) );
       }
    }
 
@@ -385,7 +338,6 @@ INTERP_QUALIFIER void TAG(copy_pv_extras)( GLcontext *ctx,
 
 
 #undef INTERP_QUALIFIER
-#undef IMPORT_QUALIFIER
 #undef GET_COLOR
 
 #undef IND
diff --git a/src/mesa/tnl_dd/t_dd_vbtmp.h b/src/mesa/tnl_dd/t_dd_vbtmp.h
index 5aad0b56222..230edf055f9 100644
--- a/src/mesa/tnl_dd/t_dd_vbtmp.h
+++ b/src/mesa/tnl_dd/t_dd_vbtmp.h
@@ -123,10 +123,10 @@ static void TAG(emit)( GLcontext *ctx,
 		       GLuint stride )
 {
    LOCALVARS
-   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
+      struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
    GLfloat (*tc0)[4], (*tc1)[4], (*fog)[4];
    GLfloat (*tc2)[4], (*tc3)[4];
-   GLubyte (*col)[4], (*spec)[4];
+   GLfloat (*col)[4], (*spec)[4];
    GLuint tc0_stride, tc1_stride, col_stride, spec_stride, fog_stride;
    GLuint tc2_stride, tc3_stride;
    GLuint tc0_size, tc1_size;
@@ -184,21 +184,16 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_RGBA) {
-      if (VB->ColorPtr[0]->Type != GL_UNSIGNED_BYTE)
-	 IMPORT_FLOAT_COLORS( ctx );
-      col = (GLubyte (*)[4])VB->ColorPtr[0]->Ptr;
-      col_stride = VB->ColorPtr[0]->StrideB;
+      col_stride = VB->ColorPtr[0]->stride;
+      col = VB->ColorPtr[0]->data;
    }
 
    if (DO_SPEC) {
       if (VB->SecondaryColorPtr[0]) {
-	 if (VB->SecondaryColorPtr[0]->Type != GL_UNSIGNED_BYTE)
-	    IMPORT_FLOAT_SPEC_COLORS( ctx );
-	 spec = (GLubyte (*)[4])VB->SecondaryColorPtr[0]->Ptr;
-	 spec_stride = VB->SecondaryColorPtr[0]->StrideB;
+	 spec_stride = VB->SecondaryColorPtr[0]->stride;
+	 spec = VB->SecondaryColorPtr[0]->data;
       } else {
-	 GLubyte tmp[4];
-	 spec = &tmp;
+	 spec = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_COLOR1];
 	 spec_stride = 0;
       }
    }
@@ -215,200 +210,124 @@ static void TAG(emit)( GLcontext *ctx,
       }
    }
 
-   if (VB->importable_data || (DO_SPEC && !spec_stride) || (DO_FOG && !fog_stride)) {
-      /* May have nonstandard strides:
-       */
-      if (start) {
-	 coord =  (GLfloat (*)[4])((GLubyte *)coord + start * coord_stride);
-	 if (DO_TEX0)
-	    tc0 =  (GLfloat (*)[4])((GLubyte *)tc0 + start * tc0_stride);
-	 if (DO_TEX1) 
-	    tc1 =  (GLfloat (*)[4])((GLubyte *)tc1 + start * tc1_stride);
-	 if (DO_TEX2) 
-	    tc2 =  (GLfloat (*)[4])((GLubyte *)tc2 + start * tc2_stride);
-	 if (DO_TEX3) 
-	    tc3 =  (GLfloat (*)[4])((GLubyte *)tc3 + start * tc3_stride);
-	 if (DO_RGBA) 
-	    STRIDE_4UB(col, start * col_stride);
-	 if (DO_SPEC)
-	    STRIDE_4UB(spec, start * spec_stride);
-	 if (DO_FOG)
-	    /*STRIDE_F(fog, start * fog_stride);*/
-	    fog =  (GLfloat (*)[4])((GLubyte *)fog + start * fog_stride);
-      }
+   /* May have nonstandard strides:
+    */
+   if (start) {
+      STRIDE_4F(coord, start * coord_stride);
+      if (DO_TEX0)
+	 STRIDE_4F(tc0, start * tc0_stride);
+      if (DO_TEX1) 
+	 STRIDE_4F(tc1, start * tc1_stride);
+      if (DO_TEX2) 
+	 STRIDE_4F(tc2, start * tc2_stride);
+      if (DO_TEX3) 
+	 STRIDE_4F(tc3, start * tc3_stride);
+      if (DO_RGBA) 
+	 STRIDE_4F(col, start * col_stride);
+      if (DO_SPEC)
+	 STRIDE_4F(spec, start * spec_stride);
+      if (DO_FOG)
+	 STRIDE_4F(fog, start * fog_stride);
+   }
 
-      for (i=start; i < end; i++, v = (VERTEX *)((GLubyte *)v + stride)) {
-	 if (DO_XYZW) {
-	    if (HAVE_HW_VIEWPORT || mask[i] == 0) {
-	       VIEWPORT_X(v->v.x, coord[0][0]);
-	       VIEWPORT_Y(v->v.y, coord[0][1]);
-	       VIEWPORT_Z(v->v.z, coord[0][2]);
-	       v->v.w = coord[0][3];
-	    }
-/*  	    fprintf(stderr, "vert %d: %.2f %.2f %.2f %.2f\n",  */
-/*  		    i, v->v.x, v->v.y, v->v.z, v->v.w); */
-	    coord =  (GLfloat (*)[4])((GLubyte *)coord +  coord_stride);
-	 }
-	 if (DO_RGBA) {
-	    if (HAVE_RGBA_COLOR) {
-	       *(GLuint *)&v->v.color = LE32_TO_CPU(*(GLuint *)&col[0]);
-	       STRIDE_4UB(col, col_stride);
-	    } else {
-	       v->v.color.blue  = col[0][2];
-	       v->v.color.green = col[0][1];
-	       v->v.color.red   = col[0][0];
-	       v->v.color.alpha = col[0][3];
-	       STRIDE_4UB(col, col_stride);
-	    }
-	 }
-	 if (DO_SPEC) {
-	    v->v.specular.red = spec[0][0];
-	    v->v.specular.green = spec[0][1];
-	    v->v.specular.blue = spec[0][2];
-	    STRIDE_4UB(spec, spec_stride);
-	 }
-	 if (DO_FOG) {
-	    v->v.specular.alpha = fog[0][0] * 255.0;
-	    /*STRIDE_F(fog, fog_stride);*/
-	    fog =  (GLfloat (*)[4])((GLubyte *)fog + fog_stride);
-	 }
-	 if (DO_TEX0) {
-	    v->v.u0 = tc0[0][0];
-	    v->v.v0 = tc0[0][1];
-	    if (DO_PTEX) {
-	       if (HAVE_PTEX_VERTICES) {
-		  if (tc0_size == 4) 
-		     v->pv.q0 = tc0[0][3];
-		  else
-		     v->pv.q0 = 1.0;
-	       } 
-	       else if (tc0_size == 4) {
-		  float rhw = 1.0 / tc0[0][3];
-		  v->v.w *= tc0[0][3];
-		  v->v.u0 *= rhw;
-		  v->v.v0 *= rhw;
-	       } 
-	    } 
-	    tc0 =  (GLfloat (*)[4])((GLubyte *)tc0 +  tc0_stride);
-	 }
-	 if (DO_TEX1) {
-	    if (DO_PTEX) {
-	       v->pv.u1 = tc1[0][0];
-	       v->pv.v1 = tc1[0][1];
-	       if (tc1_size == 4) 
-		  v->pv.q1 = tc1[0][3];
-	       else
-		  v->pv.q1 = 1.0;
-	    } 
-	    else {
-	       v->v.u1 = tc1[0][0];
-	       v->v.v1 = tc1[0][1];
-	    }
-	    tc1 =  (GLfloat (*)[4])((GLubyte *)tc1 +  tc1_stride);
-	 } 
-	 else if (DO_PTEX) {
-	    *(GLuint *)&v->pv.q1 = 0;	/* avoid culling on radeon */
+   for (i=start; i < end; i++, v = (VERTEX *)((GLubyte *)v + stride)) {
+      if (DO_XYZW) {
+	 if (HAVE_HW_VIEWPORT || mask[i] == 0) {
+	    VIEWPORT_X(v->v.x, coord[0][0]);
+	    VIEWPORT_Y(v->v.y, coord[0][1]);
+	    VIEWPORT_Z(v->v.z, coord[0][2]);
+	    v->v.w = coord[0][3];
 	 }
-	 if (DO_TEX2) {
-	    if (DO_PTEX) {
-	       v->pv.u2 = tc2[0][0];
-	       v->pv.v2 = tc2[0][1];
-	       if (tc2_size == 4) 
-		  v->pv.q2 = tc2[0][3];
+	 STRIDE_4F(coord, coord_stride);
+      }
+      if (DO_RGBA) {
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.red, col[0][0]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.green, col[0][1]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.blue, col[0][2]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.alpha, col[0][3]);
+	 STRIDE_4F(col, col_stride);
+      }
+      if (DO_SPEC) {
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.red, col[0][0]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.green, col[0][1]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.blue, col[0][2]);
+	 STRIDE_4F(spec, spec_stride);
+      }
+      if (DO_FOG) {
+	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.alpha, fog[0][0]);
+	 STRIDE_4F(fog, fog_stride);
+      }
+      if (DO_TEX0) {
+	 v->v.u0 = tc0[0][0];
+	 v->v.v0 = tc0[0][1];
+	 if (DO_PTEX) {
+	    if (HAVE_PTEX_VERTICES) {
+	       if (tc0_size == 4) 
+		  v->pv.q0 = tc0[0][3];
 	       else
-		  v->pv.q2 = 1.0;
+		  v->pv.q0 = 1.0;
 	    } 
-	    else {
-	       v->v.u2 = tc2[0][0];
-	       v->v.v2 = tc2[0][1];
-	    }
-	    tc2 =  (GLfloat (*)[4])((GLubyte *)tc2 +  tc2_stride);
-	 } 
-	 if (DO_TEX3) {
-	    if (DO_PTEX) {
-	       v->pv.u3 = tc3[0][0];
-	       v->pv.v3 = tc3[0][1];
-	       if (tc3_size == 4) 
-		  v->pv.q3 = tc3[0][3];
-	       else
-		  v->pv.q3 = 1.0;
+	    else if (tc0_size == 4) {
+	       float rhw = 1.0 / tc0[0][3];
+	       v->v.w *= tc0[0][3];
+	       v->v.u0 *= rhw;
+	       v->v.v0 *= rhw;
 	    } 
-	    else {
-	       v->v.u3 = tc3[0][0];
-	       v->v.v3 = tc3[0][1];
-	    }
-	    tc3 =  (GLfloat (*)[4])((GLubyte *)tc3 +  tc3_stride);
 	 } 
+	 STRIDE_4F(tc0, tc0_stride);
       }
-   }
-   else {
-      for (i=start; i < end; i++, v = (VERTEX *)((GLubyte *)v + stride)) {
-	 if (DO_XYZW) {
-	    if (HAVE_HW_VIEWPORT || mask[i] == 0) {
-	       VIEWPORT_X(v->v.x, coord[i][0]);
-	       VIEWPORT_Y(v->v.y, coord[i][1]);
-	       VIEWPORT_Z(v->v.z, coord[i][2]);
-	       v->v.w = coord[i][3];
-	    }
-	 }
-	 if (DO_RGBA) {
-	    if (HAVE_RGBA_COLOR) {
-	       *(GLuint *)&v->v.color = LE32_TO_CPU(*(GLuint *)&col[i]);
-	    }
-	    else {
-	       v->v.color.blue  = col[i][2];
-	       v->v.color.green = col[i][1];
-	       v->v.color.red   = col[i][0];
-	       v->v.color.alpha = col[i][3];
-	    }
-	 }
-	 if (DO_SPEC) {
-	    v->v.specular.red   = spec[i][0];
-	    v->v.specular.green = spec[i][1];
-	    v->v.specular.blue  = spec[i][2];
-	 }
-	 if (DO_FOG) {
-	    v->v.specular.alpha = fog[i][0] * 255.0;
+      if (DO_TEX1) {
+	 if (DO_PTEX) {
+	    v->pv.u1 = tc1[0][0];
+	    v->pv.v1 = tc1[0][1];
+	    if (tc1_size == 4) 
+	       v->pv.q1 = tc1[0][3];
+	    else
+	       v->pv.q1 = 1.0;
+	 } 
+	 else {
+	    v->v.u1 = tc1[0][0];
+	    v->v.v1 = tc1[0][1];
 	 }
-	 if (DO_TEX0) {
-	    v->v.u0 = tc0[i][0];
-	    v->v.v0 = tc0[i][1];
-	    if (DO_PTEX) {
-	       if (HAVE_PTEX_VERTICES) {
-		  if (tc0_size == 4) 
-		     v->pv.q0 = tc0[i][3];
-		  else
-		     v->pv.q0 = 1.0;
-
-		  v->pv.q1 = 0;	/* radeon */
-	       } 
-	       else if (tc0_size == 4) {
-		  float rhw = 1.0 / tc0[i][3];
-		  v->v.w *= tc0[i][3];
-		  v->v.u0 *= rhw;
-		  v->v.v0 *= rhw;
-	       } 
-	    } 
+	 STRIDE_4F(tc1, tc1_stride);
+      } 
+      else if (DO_PTEX) {
+	 *(GLuint *)&v->pv.q1 = 0;	/* avoid culling on radeon */
+      }
+      if (DO_TEX2) {
+	 if (DO_PTEX) {
+	    v->pv.u2 = tc2[0][0];
+	    v->pv.v2 = tc2[0][1];
+	    if (tc2_size == 4) 
+	       v->pv.q2 = tc2[0][3];
+	    else
+	       v->pv.q2 = 1.0;
+	 } 
+	 else {
+	    v->v.u2 = tc2[0][0];
+	    v->v.v2 = tc2[0][1];
 	 }
-	 if (DO_TEX1) {
-	    if (DO_PTEX) {
-	       v->pv.u1 = tc1[i][0];
-	       v->pv.v1 = tc1[i][1];
-	       if (tc1_size == 4) 
-		  v->pv.q1 = tc1[i][3];
-	       else
-		  v->pv.q1 = 1.0;
-	    } 
-	    else {
-	       v->v.u1 = tc1[i][0];
-	       v->v.v1 = tc1[i][1];
-	    }
+	 STRIDE_4F(tc2, tc2_stride);
+      } 
+      if (DO_TEX3) {
+	 if (DO_PTEX) {
+	    v->pv.u3 = tc3[0][0];
+	    v->pv.v3 = tc3[0][1];
+	    if (tc3_size == 4) 
+	       v->pv.q3 = tc3[0][3];
+	    else
+	       v->pv.q3 = 1.0;
+	 } 
+	 else {
+	    v->v.u3 = tc3[0][0];
+	    v->v.v3 = tc3[0][1];
 	 }
-      }
+	 STRIDE_4F(tc3, tc3_stride);
+      } 
    }
+
 }
 #else
-#if DO_XYZW
 
 #if HAVE_HW_DIVIDE
 #error "cannot use tiny vertices with hw perspective divide"
@@ -418,8 +337,8 @@ static void TAG(emit)( GLcontext *ctx, GLuint start, GLuint end,
 		       void *dest, GLuint stride )
 {
    LOCALVARS
-   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
-   GLubyte (*col)[4];
+      struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
+   GLfloat (*col)[4];
    GLuint col_stride;
    GLfloat (*coord)[4] = VB->NdcPtr->data;
    GLuint coord_stride = VB->NdcPtr->stride;
@@ -432,12 +351,8 @@ static void TAG(emit)( GLcontext *ctx, GLuint start, GLuint end,
 
    ASSERT(stride == 4);
 
-   if (VB->ColorPtr[0]->Type != GL_UNSIGNED_BYTE)
-      IMPORT_FLOAT_COLORS( ctx );
-
-   col = (GLubyte (*)[4])VB->ColorPtr[0]->Ptr;
-   col_stride = VB->ColorPtr[0]->StrideB;
-   ASSERT(VB->ColorPtr[0]->Type == GL_UNSIGNED_BYTE);
+   col = VB->ColorPtr[0]->data;
+   col_stride = VB->ColorPtr[0]->stride;
 
 /*     fprintf(stderr, "%s(small) importable %x\n",  */
 /*  	   __FUNCTION__, VB->importable_data); */
@@ -445,103 +360,33 @@ static void TAG(emit)( GLcontext *ctx, GLuint start, GLuint end,
    /* Pack what's left into a 4-dword vertex.  Color is in a different
     * place, and there is no 'w' coordinate.
     */
-   if (VB->importable_data) {
-      if (start) {
-	 coord =  (GLfloat (*)[4])((GLubyte *)coord + start * coord_stride);
-	 STRIDE_4UB(col, start * col_stride);
-      }
+   if (start) {
+      STRIDE_4F(coord, start * coord_stride);
+      STRIDE_4F(col, start * col_stride);
+   }
 
-      for (i=start; i < end; i++, v+=4) {
+   for (i=start; i < end; i++, v+=4) {
+      if (DO_XYZW) {
 	 if (HAVE_HW_VIEWPORT || mask[i] == 0) {
 	    VIEWPORT_X(v[0], coord[0][0]);
 	    VIEWPORT_Y(v[1], coord[0][1]);
 	    VIEWPORT_Z(v[2], coord[0][2]);
 	 }
-	 coord =  (GLfloat (*)[4])((GLubyte *)coord +  coord_stride);
-	 if (DO_RGBA) {
-	    if (HAVE_RGBA_COLOR) {
-	       *(GLuint *)&v[3] = LE32_TO_CPU(*(GLuint *)col);
-	    }
-	    else {
-	       VERTEX_COLOR *c = (VERTEX_COLOR *)&v[3];
-	       c->blue  = col[0][2];
-	       c->green = col[0][1];
-	       c->red   = col[0][0];
-	       c->alpha = col[0][3];
-	    }
-	    STRIDE_4UB( col, col_stride );
-	 }
-/*  	 fprintf(stderr, "vert %d: %.2f %.2f %.2f %x\n",  */
-/*  		 i, v[0], v[1], v[2], *(int *)&v[3]); */
+	 STRIDE_4F( coord, coord_stride );
+      }
+      if (DO_RGBA) {
+	 VERTEX_COLOR *c = (VERTEX_COLOR *)&v[3];
+	 UNCLAMPED_FLOAT_TO_UBYTE(c->red, col[0][0]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(c->green, col[0][1]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(c->blue, col[0][2]);
+	 UNCLAMPED_FLOAT_TO_UBYTE(c->alpha, col[0][3]);
+	 STRIDE_4F( col, col_stride );
       }
-   }
-   else {
-      for (i=start; i < end; i++, v+=4) {
-	 if (HAVE_HW_VIEWPORT || mask[i] == 0) {
-	    VIEWPORT_X(v[0], coord[i][0]);
-	    VIEWPORT_Y(v[1], coord[i][1]);
-	    VIEWPORT_Z(v[2], coord[i][2]);
-	 }
-	 if (DO_RGBA) {
-	    if (HAVE_RGBA_COLOR) {
-	       *(GLuint *)&v[3] = LE32_TO_CPU(*(GLuint *)&col[i]);
-	    }
-	    else {
-	       VERTEX_COLOR *c = (VERTEX_COLOR *)&v[3];
-	       c->blue  = col[i][2];
-	       c->green = col[i][1];
-	       c->red   = col[i][0];
-	       c->alpha = col[i][3];
-	    }
-	 }
 /*  	 fprintf(stderr, "vert %d: %.2f %.2f %.2f %x\n",  */
 /*  		 i, v[0], v[1], v[2], *(int *)&v[3]); */
-
-      }
    }
 }
-#else
-static void TAG(emit)( GLcontext *ctx, GLuint start, GLuint end,
-		       void *dest, GLuint stride )
-{
-   LOCALVARS
-   struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
-   GLubyte (*col)[4];
-   GLuint col_stride;
-   GLfloat *v = (GLfloat *)dest;
-   int i;
 
-   if (VB->ColorPtr[0]->Type != GL_UNSIGNED_BYTE)
-      IMPORT_FLOAT_COLORS( ctx );
-
-   col = VB->ColorPtr[0]->Ptr;
-   col_stride = VB->ColorPtr[0]->StrideB;
-
-   if (start)
-      STRIDE_4UB(col, col_stride * start);
-
-   /* Need to figure out where color is:
-    */
-   if (GET_VERTEX_FORMAT() == TINY_VERTEX_FORMAT)
-      v += 3;
-   else
-      v += 4;
-
-   for (i=start; i < end; i++, STRIDE_F(v, stride)) {
-      if (HAVE_RGBA_COLOR) {
-	 *(GLuint *)v = LE32_TO_CPU(*(GLuint *)col[0]);
-      }
-      else {
-	 VERTEX_COLOR *c = (VERTEX_COLOR *)v;
-	 c->blue  = col[0][2];
-	 c->green = col[0][1];
-	 c->red   = col[0][0];
-	 c->alpha = col[0][3];
-      }
-      STRIDE_4UB( col, col_stride );
-   }
-}
-#endif /* emit */
 #endif /* emit */
 
 #if (DO_XYZW) && (DO_RGBA)