Clarify role of MCMC_jumping_covariance in subsequent display of results

The MCMC_jumping_covariance option resets hh and thereby affects the standard errors displayed and saved for parameters and the Laplace approximation to the marginal data density. At a minimum, we should document this and add a warning in this case that the subsequent results are affected. Alternatively, we might try to move the resetting of hh to the beginning of the MCMC or condition the display and saving of the affected objects on the option not being present. The issue with moving the resetting is that this option is often specified when the actual Hessian has problems so that results based on this matrix would be meaningless as well.