Question about getFrameJacobianTimeVariation in LOCAL frame

[Linjackffy]

I found that the results of getFrameJacobianTimeVariation in LOCAL frame does not match the result from numerical differentiation. Below is the code I used.
If there are any errors in my code, please help me point them out. I would be very grateful.
If the code is correct, why would the calculation results be inconsistent?
What is the definition of the 'getFrameJacobianTimeVariation' function?

int main(int argc, char ** argv)
{
  using namespace pinocchio;

  // You should change here to set up your own URDF file or just pass it as an argument of this
  // example.
  const std::string urdf_filename =
    (argc <= 1) ? PINOCCHIO_MODEL_DIR
                    + std::string("/example-robot-data/robots/ur_description/urdf/ur5_robot.urdf")
                : argv[1];

  // Load the URDF model
  Model model;
  pinocchio::urdf::buildModel(urdf_filename, model);

  // Build a data related to model
  Data data(model);

  // Sample a random joint configuration as well as random joint velocity and acceleration
  Eigen::VectorXd q = randomConfiguration(model);
  Eigen::VectorXd v = Eigen::VectorXd::Ones(model.nv);
  Eigen::VectorXd a = Eigen::VectorXd::Ones(model.nv);
  q.setZero();
  q[0] += PI<double>();

  // Perform the forward kinematics over the kinematic tree
  forwardKinematics(model, data, q, v, a);
  FrameIndex frame_id = model.getFrameId("tool0");
  computeJointJacobians(model, data, q);
  computeJointJacobiansTimeVariation(model, data, q, v);
  updateFramePlacements(model, data);

  Eigen::MatrixXd dJ = Eigen::MatrixXd::Zero(6, model.nv);
  getFrameJacobianTimeVariation(model, data, frame_id, LOCAL, dJ);

  std::cout << "Jacobian dot : " << std::endl;
  std::cout << dJ << std::endl;
  std::cout << std::endl;

  Eigen::VectorXd q1 = q + 1e-8 * Eigen::VectorXd::Ones(q.size());
  forwardKinematics(model, data, q1);
  computeJointJacobians(model, data, q1);
  updateFramePlacements(model, data);
  Eigen::MatrixXd J1 = Eigen::MatrixXd::Zero(6, model.nv);
  getFrameJacobian(model, data, frame_id, LOCAL, J1);

  Eigen::VectorXd q2 = q - 1e-8 * Eigen::VectorXd::Ones(q.size());
  forwardKinematics(model, data, q2);
  computeJointJacobians(model, data, q2);
  updateFramePlacements(model, data);
  Eigen::MatrixXd J2 = Eigen::MatrixXd::Zero(6, model.nv);
  getFrameJacobian(model, data, frame_id, LOCAL, J2);

  Eigen::MatrixXd dJ_ = (J1 - J2) / (2e-8);

  std::cout << "Numercial Jacobian dot : " << std::endl;
  std::cout << dJ_ << std::endl;
  std::cout << std::endl;
}

Output results:

Jacobian dot : 
           0  1.87492e-12       -0.425      -1.2095      0.81725      -1.2095
           0     -0.19145     -0.19145     -0.19145      -0.2469      -0.3931
          -0     -0.09465     -0.09465     -0.09465      -0.3931  1.92487e-12
           0            1            1            1            3            0
           0  9.79328e-12  9.79328e-12  9.79328e-12  4.79571e-23  9.79328e-12
          -0 -4.79541e-23 -4.79541e-23 -4.79541e-23  9.79328e-12 -4.79541e-23

Numercial Jacobian dot : 
     0.81725      -2.0595      -0.7845            0  6.93889e-10   4.5172e-40
     -0.7658     -0.17695     -0.17695     -0.17695       0.0823  1.20719e-35
     -0.3931     -0.09465     -0.09465     -0.09465  4.99773e-10  2.46519e-24
           4            1            1            1           -1 -2.38256e-16
           0 -4.03897e-20 -4.03897e-20 -4.03897e-20            0 -8.07794e-20
 9.79328e-12            0            0            0  4.03897e-20            0

[jcarpent]

You're correct, one term is missing.
Yet, the result Jdot * v is correct ;)

[Linjackffy]

With the latest devel branch, this issue has been resolved. Thanks for your help!