I think there are no experiments in the lab or a direct measurement of this. Anyway, if the light is blended during an eclipse, the that light should have accelerated the Sun just a tiny tiny ... tiny tiny bit in the opposite direction.