Code review comment for lp:~nickpapior/siesta/4.1-wfs

I agree with your tSpin and print_spin changes. They do make the code more clear. Thanks.

The new tSpin variable (equivalent to Ncol.or.SO), and other enhancements, can maybe be defined in the 4.2 branch.

Regarding the output files headings:

  In bands and ioeig, no_u*2 refers to the number of states, and 1 to the
  number of spin blocks (=1 since there is no spin distinction).

  In the wavefunction files, no_u refers to the number of basis orbitals, and
  4 to the number of expansions of length no_u needed at each band (4 for a complex
  spinor). The number of states in the file is given by nk and the following records
  for each k-point.

One possible source of fragility is that the "spinor_dim second-dimension" trick in ioeig and bands might not work when 'neigwanted' is not no_u. This could be also handled more robustly in 4.2, after the merge, and back-ported if deemed necessary for some use cases.