What if you had to compile a language that is actively trying to make compilation impossible? What if every program had to be viewed as a maleficent adversary trying to break any attempt to optimize its code? Well, we would say that you are probably trying to compile R, the programming language of choice for statistics. This talk feature tall tales of trying, and often failing, to compile R. We will explain what makes R adversarial and why it got that way. We will discuss steps that language implementers take to protect themselves, and conclude with ideas on how to smooth the sharp edges of the language.