I am working on a glsl shader program as part of a plugin that runs inside a "closed source" application. The application (maya) is written using opengl 2.1, but our graphics cards support opengl/glsl 4.1 and I want to use tessellation and geometry shaders in my program. The application sets up the opengl viewport and the traditional model/view matrix stack and I do not have control over that part of the code.
My pass-through vertex shader uses GLSL 1.2 and works fine:
// GLSL VERTEX SHADER
#version 120
void main ()
{
gl_Position = gl_ModelViewProjectionMatrix * gl_Vertex;
}
My pass-through geomery shader uses glsl 4.1 and also works fine inside the application:
// GLSL GEOMETRY SHADER
#version 410
layout (lines_adjacency) in;
layout (line_strip, max_vertices = 2) out;
void main ()
{
gl_Position = gl_in[1].gl_Position;
EmitVertex();
gl_Position = gl_in[2].gl_Position;
EmitVertex();
EndPrimitive();
}
But this is just a pass-through test. In my real geometry shader I need to do some calculations in world-space but the geometry shader points are in view-space. My question is: can I access the gl_ModelViewProjectionMatrix inside a 4.1 geometry shader? I understand that the traditional matrix stack has been deprecated in glsl 4.1 in favor of uniform variables, but I cannot change the application. I cannot use glsl 1.2 in my geometry shader because I need the lines_adjacency input type. Do I need to copy the matrix into a uniform variable in my plugin's C++ source? Or is there a "back door" to get at it directly from glsl 4.1? Or something else I'm not thinking of?
