As far as I can see there's nothing wrong with DiracOrder.
What's happening is just that (see http://www.feyncalc.org/Polarization/index.html) the contraction of a momentum frou-vector and the corresponding polarization vector is automatically set to zero.
That is, transversality is assumed. If you don't want that, use some other name than Polarization[...] for the polarization vector.
Hope this helps.
This archive was generated by hypermail 2b29 : 10/23/17-01:20:01 PM Z CEST