Reorder states and probability tables

[Manuel Luque]

[This is an old issue. I copy and paste here the messages written by Manuel Luque and Javier Díez several months ago]

mluque: When the states of a node are reordered the numerical values in its conditional probability table are kept, but they aren't reordered correctly.

fjdiez: The change should be made not only for probability tables, but also for utility tables, ICI models, etc. However, other types of potentials, such as Uniform and Tree/ADD, do not need it. We should consider carefully the implications of this change.

[Francisco Javier Díez]

It has been solved for probability and utility tables. For other types of potentials an analogous change is not desirable, at least in some cases. So we mark the issue as resolved.