I don't understand how can datenwolf`s index generation can be correct. But still I find his solution rather clear. This is what I get after some thinking:
inline void push_indices(vector<GLushort>& indices, int sectors, int r, int s) {
int curRow = r * sectors;
int nextRow = (r+1) * sectors;
indices.push_back(curRow + s);
indices.push_back(nextRow + s);
indices.push_back(nextRow + (s+1));
indices.push_back(curRow + s);
indices.push_back(nextRow + (s+1));
indices.push_back(curRow + (s+1));
}
void createSphere(vector<vec3>& vertices, vector<GLushort>& indices, vector<vec2>& texcoords,
float radius, unsigned int rings, unsigned int sectors)
{
float const R = 1./(float)(rings-1);
float const S = 1./(float)(sectors-1);
for(int r = 0; r < rings; ++r) {
for(int s = 0; s < sectors; ++s) {
float const y = sin( -M_PI_2 + M_PI * r * R );
float const x = cos(2*M_PI * s * S) * sin( M_PI * r * R );
float const z = sin(2*M_PI * s * S) * sin( M_PI * r * R );
texcoords.push_back(vec2(s*S, r*R));
vertices.push_back(vec3(x,y,z) * radius);
push_indices(indices, sectors, r, s);
}
}
}