Comment on "Modified photon equation of motion as a test for the principle of equivalence"

In a recent paper, a modification of the geodesic equation was proposed for spinning photons containing a spin-curvature coupling term. The difference in arrival times of opposite circular polarizations starting simultaneously from a source was computed, obtaining a result linear in the coupling parameter. It is pointed out here that this linear term violates causality and, more generally, Fermat's principle, implying calculational errors. Even if these are corrected, there is a violation of covariance in the way the photon spin was introduced. Rectifying this makes the effect computed vanish entirely.