In some rare cases, if one starts with e.g. an antiferromagnetic configuration, and after the first cycle ends up in a different spin configuration, the current the relabelling would fail. It is possible to fix it, but then we raise the question whether to continue on with the calculation, or stop it and report an exit code that the magnetic configuration is probably different from the starting one.

Example:

1. Iteration: start with e.g. Fe0 and Fe1, having starting magnetization +1 and -1 respectively
2. Iteration: the scf leads to Fe0 and Fe1 having 0 magnetic moments --> hp will set the label spin to 0, and this will make the relabelling to crash.

How to fix:

• Simply catch the excpetion, either within the relabelling (maybe better?) or the self-consistent cycle.