direction of normal raise of results for triangle elements is wrong
I have such a terrible suspicion that in gmsh.view.addModelData function (Julia) the normal vector of the triangle (2D) elements (in postproc view) is inverted compared to quad elements. In example x3.jl addListData function was used and there the sequence of the nodes is prescribed (in x3.jl). If I change the sequence there (from counterclockwise to clockwise), the normal of the element (in postproc view) will change. That why I think that there is a bug in gmsh.view.addModelData function. For quad element it works fine, but for triangles not.