The calculation can fail due to the structure not being sorted

[JPchico]

Currently if the StructureData node does not have its sites sorted in such a way that atoms are grouped via type this might result in VASP itself crashing. this happens in large supercells

The reason is that if the atoms are not sorted when the program is trying to read the atomic species a check in how long the line can be is reached causing a conflict between the number of atoms and sites.

One could provide a warning for this, in which one could try to detect if the structure is sorted, perhaps this is something for parsevap?

Another thing that could be done is providing a calcfunction that takes the structure and sorts it so that VASP is happy. Then it would be to the user to determine whether or not to use this functionality.