On the apparent failure of the topological theory of phase transitions