'Reject' definition.

[whatwg]

'Reject' definition.

[bzbarsky]

Oh, this went away with platform array objects, looks like. That's bad.

[bzbarsky]

In particular, in #52

[bzbarsky]

Oh, but more to the point, we don't need this thing because in ES6 the contract for [[DefineOwnProperty]] doesn't involve throwing at all. You just return false and the caller throws if needed.

@heycam should we just update all out [[DefineOwnProperty]] to the ES6 definition?

[heycam]

Yes I think we should update. And to me it looks like all [[DefineOwnProperty]] calls in Web IDL currently can just ignore its return value. Do you agree? Apart from http://heycam.github.io/webidl/#defineownproperty which needs to just return it.

[bzbarsky]

@heycam The calls in dictionary and sequence idl-to-es conversion should never return false. Same thing in the unscopeables setup and serializer bits, So ignoring the return value there makes sense.

The call for [Replaceable] attribute setters, however, I think should probably throw if false is returned. Testing this in browsers is a bit of a pain because you can't freeze a Window effectively, but this is a testcase:

var str = "Object.freeze(self); self.console = 5; postMessage(typeof console);"
var blob = new Blob([str]);
var url = URL.createObjectURL(blob);
var w = new Worker(url);
w.onmessage = function (e) {
  alert("OK: " + e.data);
}
w.onerror = function (e) {
  alert("Error: " + e.message);
}

Looks to me like this throws in Firefox but not Chrome (which silently ignores the set instead)... Chrome doesn't throw with use strict there either, of course. In any case, I think we basically want the behavior here to be like Object.defineProperty, which would throw in this situation, so throwing makes sense to me.

[ylafon]

Note that the this was fixed by 8c9b828