Splitting Matsubara into two or four different meshes: yes or no?

[egull]

The Matsubara mesh right now has two bools: negative/poisitve or only positive, and boson or fermion. Should we split these into two or four meshes?

   Suggestion: split this mesh into for disjoint meshes, one for fermions, one for bosons, each for specifying both positive and negative or only positive frequencies.

  Andrey: [Q] There seem to be a number of conditionals - this is error-susceptible for future implementations.