There is a serious bug!

The structure file must currently have the first transformation to be unity, otherwise the rotations do not work properly.
Improve the rotations such that this requirement is not necessary!!!  The error was reproduced on Cerium.

The reason for the problem is that we add a rotation "rotij" which is supposed to transform from the first atom to all the rest.
If there is only one atom, it seems that this rotation is the first from the structure file. This is not acceptable.
You should understand what is roij in general, and how should be used in our code "l2main.f".

One more bug:
When the correlated atom is not the first atom and one is performing spin polarized calculation, there seems to be a bug.
Fix: put correlated atoms first.



DMFT0: now it takes NOE from scf2 file. It would be better to take from case.in2 file, just like dmft2



1) Edc.dat should be removed before the job starts
2) dmft1.error.* should be removed before the job starts

