Evol: NNLO Polarized

[giacomomagni]

While doing the benchmark of NNLO polarized evolution with Eko,
I found two things which are not fully consistent:

• P_{ns,s} is not implemented here, but it's available from this source ref eq 23, 25.
  This should have a minor effect on the Valence distribution.
• As far as I understand, VFNS evolution is not possible since polarized matching conditions are not implemented.

[vbertone]

Hello Giacomo, yes, you are correct on both counts.
I wasn't indeed aware that P_{ns,s} had been calculated. Thanks for pointing it out.
I will try to have it implemented as soon as possible.