1.2-31: Changed a bug in jointprior thanks to Jean-Baptiste DENIS. A non-uniform prior would be multiplied together in a wrong fashion. 1.2-31: Changed a bug in autosearch and heuristic thanks to Peter van de Ven. The output network was not necessarily the one with highest loglikelihood due to a wrong sorting of the scores. 1.2-32: The bug in jointprior was reverted. Apparently not a bug and it would make the demo(ksl) go down. 1.2-34: Minor changes due to new R version