Use same (initial) x in get_nlp_info and the initial eval_jac_g/eval_h
calls.
Previously, a zero vector was used for the latter evaluations, which can
result in a different number of non-zero matrix entries, leading IPOPT
to allocate too little.