Problem with computing Hessian in mixed mode (math library)

[mirkhosro]

I found this thread in the discussions which is the most similar to the problem I have:

Matrix transformations of parameters (Stan math library)

I am trying to build a model in which my parameter vector is transformed in various ways before calculating the likelihood. Eventually, I will want to calculate the gradient and Hessian of the likelihood. One or more of the transformations of (parts of) the input vector will involve matrix-matrix or matrix-vector multiplication, where the matrices and vectors involved will either contain doubles, or (transformations of) parameter elements. As a simple example, here is some code which calculates…

I wonder if the recommended workaround is still the same as the ones suggested there.