Why does arguments.callee hurt performance?

Question

The alternative to using arguments.callee is to simply name the function, as such:

// Snippet 1, using arguments.callee:
(function(){
    document.body.innerHTML += 'qwe';
    setTimeout(arguments.callee, 1);
})()

// Snippet 2, using naming:
(function f(){
    document.body.innerHTML += 'qwe';
    setTimeout(f, 1);
})()

What is stopping the JavaScript engine/compiler from automatically optimizing Snippet 1 into Snippet 2 above? Is there some inherent limitation existing?

• MDN's argument basically boils down to:

  ..arguments.callee substantially hinders optimizations like inlining functions, because it must be made possible to provide a reference to the un-inlined function if arguments.callee is accessed.

  However, if we manually name the function and call it via it's name, we are already providing a reference to the function, and thus "hindering optimizations like inlining functions".

• Olliej's argument is basically "inlining and tail recursion [becomes] impossible in the general case [whereas it is possible if naming is used]". However no evidence is provided and it seems like a handwave (Also see the comments below his post).

Why is function-naming more performant than using arguments.callee?

Answer 1

By my reading, MDN is not saying "it's hard to provide a reference to the un-inlined function"; rather, it's saying "it's hard to ensure that, when you access arguments.callee, you get a reference to the un-inlined function".

(This is only half an answer to your question, because it's still not clear to me why arguments.callee should be so much harder to keep up-to-date than the rest of arguments, or than this. So the argument, as it stands, is not obviously compelling. But that's what I think MDN is saying.)