Fix test under Octave, by changing the value of mode_compute

With the former optimizer, the ML estimation under Octave would lead to a point
with a non-positive covariance matrix for shocks; in turn this would be an
unsuitable starting point for the subsequent Bayesian estimation

tests/measurement_errors/fs2000_corr_me_ml_mcmc/fs2000_corr_ME.mod

@@ -104,7 +104,7 @@ corr gp_obs, gy_obs,0;
 end;
 
 options_.TeX=1;
-estimation(mode_compute=9,order=1,datafile=fsdat_simul,mode_check,smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
+estimation(order=1,datafile=fsdat_simul,mode_check,smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;