[Wien] in1new problem (addition)

[Stefaan Cottenier]

In addition to the previous post: I meanwhile tested that for a LuSb in 
the same geometry, in1new and spin-orbit work without problem. For Lu, 
the 5d are in the core. The linearization energy for Sb-5p is found by 
in1new as the weighted average of the two types of spin-orbit 
eigenvalues, as should be. Conclusion: the problem for LaSb is really 
due to the presence of La-5d in the same energy region.

Stefaan