useful normals for ofMesh::plane

I tried extruding an ofMesh::plane as returned by ofPlanePrimitive, then calculating the normals, but I think the winding alternates from triangle to triangle.

Here's what it looks like with ofPlanePrimitive:

And here's with custom code for generating the mesh:

Here are the two relevant excerpts from the code:

ofPlanePrimitive plane;
plane.set(w, h, w, h);
mesh = plane.getMesh();
mesh.clearNormals();

glm::vec3 center(w/2, h/2, 0);
for(int y = 0; y < h; y++) {
    for(int x = 0; x < w; x++) {
        mesh.addVertex(glm::vec3(x, y, 0) - center);
    }
}

for(int y = 0; y < h-1; y++) {
    for(int x = 0; x < w-1; x++) {
        int nw = y * w + x;
        int ne = nw + 1;
        int sw = nw + w;
        int se = sw + 1;
        mesh.addIndex(nw);
        mesh.addIndex(ne);
        mesh.addIndex(sw);
        mesh.addIndex(sw);
        mesh.addIndex(ne);
        mesh.addIndex(se);
    }
}

And normal calculation:

void addNormals(ofMesh & mesh){
    auto& vertices = mesh.getVertices();
    auto& normals = mesh.getNormals();
    auto& indices = mesh.getIndices();
    
    for(auto& v : vertices) {
        mesh.addNormal({});
    }
    
    for(int i = 0; i < indices.size(); i+=3) {
        int ia = indices[i];
        int ib = indices[i+1];
        int ic = indices[i+2];
        
        auto e1 = vertices[ia] - vertices[ib];
        auto e2 = vertices[ic] - vertices[ib];
        auto no = glm::cross(e1, e2);
        
        normals[ia] += no;
        normals[ib] += no;
        normals[ic] += no;
    }

    for(auto& normal : mesh.getNormals()) {
        normal = glm::normalize(normal);
    }
}

And here is a zoomed version of the ofPlanePrimitive one to show the triangulation:

If this is inherent to triangle strips, maybe there should be an option for building an OF_PRIMITIVE_TRIANGLES mesh instead.

Start by inspecting the ofPlanePrimitive mesh generation and ofMesh plane/index handling, then reproduce the reported normal calculation with the snippets in the issue. Compare triangle winding and resulting normals with the custom mesh; done means the cause is confirmed and the expected consistent-normal behavior or triangle-mesh option is covered by a regression test.