#136 Link/join polarizations + spectral fitting

[André Offringa]

Now that IDG is getting more and more into production, and we have the linked polarizations option, it has become a common case to want to link on Stokes I and simultaneously fit. It's a bit awkward what to do with the other polarizations (leave unconstrained? fit independently? make options?), but a solution should be found, because we need this for I.