The p-process revisited