[patch review] add gl_vertex_program_point_size to state_pscale

[Erik Inge Bolsø]

Hello!

I'd like a review of this patch - does the approach look reasonable? (Not 
my code, but it fixes a bug I reported.)

"The fires and smoke aren't fading, they're just not being sized properly. 
The game uses point sprites for these effects. No matter what distance, 
the pointsize is being set to 1."

It fixes http://bugs.winehq.org/show_bug.cgi?id=8826 - I'm not sure if 
it breaks other apps yet.

Patch from Cody Hamilton  :

*** state.c
--- dlls/wined3d/state.c	2009-03-05 00:42:57.000000000 +1100
***************
*** 1448,1453 ****
--- 1448,1461 ----
          att[1] = B.f / scaleFactor;
          att[2] = C.f / scaleFactor;
      }
+     
+ 	if(stateblock->wineD3DDevice->shader_backend == &glsl_shader_backend) {
+ 		glEnable(GL_VERTEX_PROGRAM_POINT_SIZE);
+ 		checkGLcall("glEnable(GL_VERTEX_PROGRAM_POINT_SIZE)");
+ 	} else {
+ 		glDisable(GL_VERTEX_PROGRAM_POINT_SIZE);
+ 		checkGLcall("glDisable(GL_VERTEX_PROGRAM_POINT_SIZE)");
+ 	}

if(GL_SUPPORT(ARB_POINT_PARAMETERS)) {
          GL_EXTCALL(glPointParameterfvARB)(GL_POINT_DISTANCE_ATTENUATION_ARB, att);

--
-erik	http://useofwords.blogspot.com/

[Henri Verbeet]

2009/3/12 Erik Inge Bolsø :
> Hello!
>
> I'd like a review of this patch - does the approach look reasonable? (Not
> my code, but it fixes a bug I reported.)

Not in this form, although enabling GL_VERTEX_PROGRAM_POINT_SIZE is
something we should do. It needs to go either into
shader_glsl_select() or into a shader backend specific context
initialization function (this would need to be added first), depending
on how D3D interprets pointsize for shaders that don't write to oPts.