Eigen vs Cholesky decomposition

[mariodosreis]

There seems to be some inconsistent between how the Eigen and Cholesky decompositions of correlation matrices are implement. Results are correct, but the method used is weird. Need to check the write.morpho and other functions that use the decomposition, and standardise. Perhaps we should just keep the Eigen decomposition. For example, for post hoc ancestral sequence reconstruction, it is important to use the same decomposition used when writing the morphological alignment, otherwise the reconstruction may be wrong (need to check this). Using a single method for the decomposition would avoid this problem.