I think this is just a symptom of the feature not being shipped, and being behind a flag.

I'll send an intent-to-ship email tomorrow

> I think this is just a symptom of the feature not being shipped, and being behind a flag.

No.  This issue means enabling the feature breaks interoperability.
The test passes on Firefox and Edge.

Right, so there are a few confusing things here:

The test name sounds like it's referring to https://github.com/whatwg/html/pull/638, in particular the changes to internal methods for these cross origin accesses --- but the actual content of the test is not verifying this at all, due to not operating on a different browsing context at all. So I'm fairly sure the test itself is wrong.

Other than that, it looks like other browsers just don't implement the feature yet at all, which would explain why they pass an apparently incorrect test

I've filed https://github.com/w3c/web-platform-tests/issues/2664 to see if this is a test bug or not --- in the mean time, I doubt it's actually breaking any content anywhere, but if it is, lets remove the stuff adding @@iterator to window.

I don't think cross-origin access matters here.  Why does window have @@iterator?

It's in the WebIDL spec, https://heycam.github.io/webidl/#es-iterators

In particular: "window" has a "length" attribute, and an indexed getter --- thus, per WebIDL, it has @@iterator.

The test itself (if you look at the file name) is entirely related to accessing different browsing contexts, and all of the other tests in the directory excercise this. So it seems like this was probably intended to be about the restrictions on ES internal methods when talking to browsing contexts from a different domain, but I'd like to verify this with Ms2ger

After talking to bz about it, it sounds like there was a spec change to remove this feature from window, which Domenic should know something about.

So, I suppose we can go ahead and remove it from window.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/02a0abbbdb24780320afc40d02c3114b1045b8e0

commit 02a0abbbdb24780320afc40d02c3114b1045b8e0
Author: caitpotter88 <caitpotter88@gmail.com>
Date: Wed Mar 16 16:14:02 2016

[bindings] remove @@iterator from Window

Window is now exempt from the WebIDL policy granting
@@iterator to Array-like objects, as of a recent change to the Window
specification (https://html.spec.whatwg.org/#window) --- due to the
indexed getter being removed, and handled in WindowProxy instead (https://github.com/whatwg/html/commit/acae3df652b382e9f4f1d1b4dc7e08e2b00df821)

BUG= 593174 ,  538558 
LOG=N
R=yukishiino@chromium.org, haraken@chromium.org

Review URL: https://codereview.chromium.org/1778773002

Cr-Commit-Position: refs/heads/master@{#381464}

[delete] https://crrev.com/46388e9ff6e06a2d24edd029df4a9149e70795c1/third_party/WebKit/LayoutTests/imported/web-platform-tests/html/browsers/the-window-object/accessing-other-browsing-contexts/iterator-expected.txt
[modify] https://crrev.com/02a0abbbdb24780320afc40d02c3114b1045b8e0/third_party/WebKit/Source/bindings/templates/interface_base.cpp
[modify] https://crrev.com/02a0abbbdb24780320afc40d02c3114b1045b8e0/third_party/WebKit/Source/bindings/tests/results/core/V8TestIntegerIndexedGlobal.cpp
[modify] https://crrev.com/02a0abbbdb24780320afc40d02c3114b1045b8e0/third_party/WebKit/Source/bindings/tests/results/core/V8TestIntegerIndexedPrimaryGlobal.cpp