Lagrange multipliers and Virial theorem in default mode

[jongrumer]

By default, DBSR_HF has rotations turned off - which can lead to a wrong solution (since Lagrange multipliers are excluded). This can be seen from a poor agreement in the Virial theorem, which probes whether or not the obtained solution is stationary.

@cffischer suggests to put rotation=1 and debug=1 turned on by default - so that Lagrange multipliers are used, and the Virial theorem printed to the .log file, so that the user can always see if the solution is good. This will take some more time, computationally, but should be more failsafe.

Ground state of Cu I: 4d10 4s is a classic example where Lagrange multipliers are needed to obtain a correct solution.