what parameters does the function texture() takes?

Question

float texture(gsampler2DArrayShadow sampler, vec4 P, [float bias]);

Look at this function in OpenGL ES (shader lang). I do not understand the difference between gsampler2DArrayShadow and sampler2DArrayShadow. Can you explain this? I read that 'g' can mean nothing or i or u. But what then is this type?

Also, Does '[float bias]' mean that we have 2 functions: with this parameter and without?

Answer 1

Anytime you see a variable type in GLSL prefixed with g in a function prototype, that is a shorthand convention that means there is an overload for every type of data. A function that accepts gvec, for instance, means it has an overload for ivec, uvec, vec, dvec, bvec and so on.

Desktop GLSL has support for integer samplers in addition to fixed-/floating-point so you will see a lot of functions defined using gsampler... instead of sampler..., that means the function has isampler... and usampler... overloads in addition to the more traditional variety.

---

As for [float bias], that is used to control mipmap LOD bias. When you do not supply a value to that parameter, GL computes the bias itself.

Interestingly because of the way mipmap LODs are computed, only the fragment shader variety of the texture lookups support automatic mipmap selection (it has to do with the per-fragment derivative calculation). To use mipmapping in the vertex shader, you have to explicitly select the LOD with textureLod (...) or supply your own partial derivative using textureGrad (...).