EKF: Jacobian Matrix for the state estimate error covariance matrix (P) prediction

[afonsomm]

https://github.com/rlabbe/filterpy/blob/3b51149ebcff0401ff1e10bf08ffca7b6bbc4a33/filterpy/kalman/EKF.py#L353-L371

Shouldn't the Jacobian matrix of the system's dynamics be used in the prediction of (P) (line 367), instead of the transition matrix (F)?