The order is undefined -> "I, the compiler, declare this branch can't happen" -> delete. The middle step is valid because "undefined behavior" permits the compiler to make that declaration, then act on it. If you don't want it to do that, use defined behaviors only, which is a great deal easier said than done. Partially because of how hard it is to avoid it in your own code, and partially because it is shot through all the other library code (which as pcwalton points out, courtesy of aggressive inlining, is also your code).

I was skeptical about all this about six months ago myself, but the continuous stream of articles on this topic, plus the spectacular and highly educational failure of Friendly C (and not just that it failed, but why it failed, which is why I posted that exact link) has satisfied me. It is also part of why I've stepped up my own anti-C rhetoric since I've been so convinced... as bad as I thought C was, it really is, no sarcasm, not merely "hating", even worse than I thought. I am honestly scared to use it professionally and pretty much refuse to touch it without good static analysis support.