Alternative fix for #3508

[Rich-Harris]

This is an alternative to #3809 (fixes #3508). It feels like a bit of a hack, but if we replace the array passed to the end of init with an object whose values are zeroes...

-init(this, options, instance, create_fragment, safe_not_equal, ["a", "b"]);
+init(this, options, instance, create_fragment, safe_not_equal, {a:0, b:0});

...then we can add renamed properties to the same object...

init(this, options, instance, create_fragment, safe_not_equal, {a:0, b:0, c:"d"});

...and handle aliased exports while adding slightly less code:

export function bind(component, name, callback) {
  if (name in component.$$.props) {
    name = component.$$.props[name] || name;
    component.$$.bound[name] = callback;
    callback(component.$$.ctx[name]);
  }
}

I don't feel great about it, but it works. @Conduitry thoughts?

[Conduitry]

Linting doesn't like using Object.prototype.hasOwnProperty apparently - and probably for good reason, as I think this would get messed up with a prop called hasOwnProperty. What I had meant before was actually using Object.hasOwnProperty(component.$$.props, name).

[Rich-Harris]

wait, when did hasOwnProperty become a static method on Object? It works in Chrome, but does it work everywhere? I always used to do Object.prototype.hasOwnProperty.call(obj, prop)

[Conduitry]

Apparently I'm going crazy. I just found this SO question, asked a mere three days ago. It sounds like Object.hasOwnProperty doesn't really exist, but, because of javascript mysteries, you can call it, but it doesn't actually do what I expected. Sorry about that! Object.prototype.hasOwnProperty.call(component.$$.props, name) sounds safest.

Edit: Or ({}).hasOwnProperty.call(...) to save bytes, I guess. I don't know whether that's a performance hit.

[Rich-Harris]

Turns out we're using hasOwnProperty somewhere already so I turned it into a has_prop helper