test_dh17.jl has a commented out section in which it calls dh17! with BigFloat precision to compute the Jacobian. The returned jacobian has an error in which each data point is off by a factor of order unity; it looks like there is a problem with the passing of a BigFloat array to the subroutine and modifying it within. This was solved by adding an output to dh17!, but in general this isn't a workable solution.
test_dh17.jl has a commented out section in which it calls dh17! with BigFloat precision to compute the Jacobian. The returned jacobian has an error in which each data point is off by a factor of order unity; it looks like there is a problem with the passing of a BigFloat array to the subroutine and modifying it within. This was solved by adding an output to dh17!, but in general this isn't a workable solution.