I recently bought the third edition of the orange OpenGL Shader Language and found that a lot had changed since the OpenGL Superbible was last updated. One of the biggest changes was that global uniforms such as gl_Vertex and gl_Normal were deprecated, and instead we were to use the 'in' keyword to show we are expecting input. One problem: the Vertices seem fine, but the normals given to me seem incorrect. If i use gl_Normal my shader works fine, but if i use the following line: "in vec3 MCnormal;", the normal never seems to change. I'm rendering a glutSolidTorus, so the normals should be correct, and as I said before, if i just reference "gl_Normal" i get results. Any idea why this is? I'm still learning the changes surrounding OpenGL 3.2, so perhaps I'm missing something in my code? Besides passing in the ModelViewMatrix, ModelViewProjectionMatrix and NormalMatrix, everything else seems to be basically the same. Vertex Shader:

#version 130

in vec4 MCvertex;
in vec3 MCnormal;

uniform mat4 MVMatrix;
uniform mat4 MVPMatrix;
uniform mat4 NormalMatrix;

out float LightIntensity;

void main()
{
  float diffuse = 0.7;
  vec3 lightPos = vec3(0,10,0);
  mat3 NMat = mat3(NormalMatrix);
  
  vec3 position = vec3(MCvertex * MVMatrix);
  
  vec3 L = normalize(lightPos - position);
  
  vec3 N = normalize(NMat * normalize(MCnormal));
  
  LightIntensity = diffuse * max(dot(N, L), 0.0);
  
  gl_Position = MCvertex * MVMatrix * MVPMatrix;
}

Fragment Shader:

#version 130

in float LightIntensity;

out vec4 FragColor;

void main()
{
  FragColor = vec4(LightIntensity, LightIntensity, LightIntensity, 1.0);
}

Sorry, I meant attributes. I was thinking of all of the matrices (ModelView, ModelViewProjection etc...)

I looked more into this issue, and found many people are sending the vertex and normal data in through buffers, however I am using the glutSolidTorus() function, which renders the object for me. I don't have access to its normals or vertices, so I don't think I can use glVertexAttribPointer().

Any suggestions? Do functions like glutSolidTorus and glutSolidSphere no longer operate the way they are supposed to because of this new way of passing information to the shaders?

So the standards for glsl 130 more or less disallow the use of glutSolid* calls. Does that mean that we no longer have access to simple primitives? I feel this significantly complicates things.

Does this mean that all glNormal calls are now deprecated? Can we no longer use glBegin and glEnd with simple vertex, normal, and color calls to define simple polygons?

That is understandable. Any idea if the glut libraries will ever be updated to follow this new standard?

Also, through all of this I have found a profound lack of clear tutorials or documentation on glsl 130 or higher. most of my problems stemmed from the book i was referencing (OpenGL Shading Language, 3rd Edition,) which doesn't have ANY source code beyond the vertex and fragment shaders to clear things up. It then took a lot of work to find out how to bring my code into the 'core' profile.

I would just like to see how other people are handling the new standards. Much of my work has been touch and go so far.

Quote:Original post by Pirosan
Any idea if the glut libraries will ever be updated to follow this new standard?

Freeglut, maybe. GLUT itself was abandoned 12 years ago.

In general, however, basic primitives are not useful for anything other than brief tests - and those are simple enough to handle with a model loaded from OBJ/similar format.