Heavy crash when using RMS maps which have missing pixels

[hermda02]

Line 374 of comm_output_mod does not seem to take into account masked pixels when computing chisq_map_eff. Does anyone know if the way this is defined is vital or if we can just loop over the pixels?

[dncnwtts]

I wrote this bit with the log(sigma0) when I forgot about the rms maps for incomplete sky. I think this can be removed completely.