What is the most efficient way to add a Eigen::DiagonalMatrix to an Eigen::SparseMatrix

Question

I am surprised to find out that it's not possible to simply add an Eigen::DiagonalMatrix to an Eigen::SparseMatrix:

#include <Eigen/Core>
#include <Eigen/Sparse>

int main()
{
  Eigen::Matrix4d M;
  M<<1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16;
  Eigen::SparseMatrix<double> S = M.sparseView();
  Eigen::DiagonalMatrix<double,Eigen::Dynamic> D = 
    Eigen::VectorXd::LinSpaced(4,1,4).asDiagonal();
  Eigen::SparseMatrix<double> T = D+S;
}

will not compile:

 error: invalid operands to binary expression ('Eigen::DiagonalMatrix<double, Eigen::Dynamic>' and 'Eigen::SparseMatrix<double>')
  Eigen::SparseMatrix<double> T = D+S;

It seems in Eigen 3.3 I can use:

  T.diagonal() = D.diagonal() + T.diagonal();

But this doesn't compile for the older (more stable) Eigen 3.2.*:

error: no viable overloaded '='
  T.diagonal() = D.diagonal() + T.diagonal();

Do I really have to iterate over the entries and insert them one by one?

 Eigen::SparseMatrix<double> T = S;
 for(int i = 0;i<T.rows();i++) T.coeffRef(i,i) += D.diagonal()(i);

Update:

Looks like you can also do:

Eigen::SparseMatrix<double> T = Eigen::SparseMatrix<double>(D)+S;

I'm guessing this incurs unnecessary copying.

Panic