On page 35 of Sidney Coleman's Classical Electron Theory from a Modern Standpoint, he writes:
If we remember that in Maxwellian dynamics, the radiation field is given by the difference of the advanced and retarded fields, it is easy to see that the electron radiates.
The figure shows the world line of an electron undergoing hyperbolic acceleration. The past light cone from the point P does not interact with the electron worldline, so the retarded potential vanishes; the future light cone intersects it once, so the advanced potential doesn't. Thus there is a radiation field at P.
The retarded potential at P is undefined, so how is Coleman able to conclude that it vanishes?