Remove noopener from tokenizedFeatures

[annevk]

Apparently CSSOM needs to react differently if it's there.

Helps with #1902.

---

/infrastructure.html ( diff )
/window-object.html ( diff )

[annevk]

So given #1902 (comment) I guess with this fix it would still be good to have tests and implementation bugs. Do you know if we have tests?

[zcorpan]

I think CSSOM View still needs a fix for https://www.w3.org/Bugs/Public/show_bug.cgi?id=25722 / #2464 for this change to make a difference.

I see these tests for noopener, but I think the tests we have don't check for "tab" vs "window".

https://github.com/w3c/web-platform-tests/blob/master/html/browsers/the-window-object/window-open-noopener.html
https://github.com/w3c/web-platform-tests/blob/master/html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-tokenization-noopener.html

[annevk]

I've made this account for noreferrer as well now. And I landed some tests for BarProp stuff, but unfortunately those don't seem to do the trick in Chrome.

[annevk]

I think I'd be okay with landing this given that there's now some tests, even though they don't quite show the issue in Chrome yet (they will after a bug is fixed).

@bzbarsky what do you think, does this improve the status quo even if the CSSOM bits aren't done? Should there be another test that compares non-BarProp aspects of the windows, e.g., viewport width or height?

[bzbarsky]

I think this does improve things even without the CSSOM bits.

Doing a test that compares innerWidth/innerHeight and outerWidth/outerHeight with the "known good" window does seem like a good idea.

[annevk]

@zcorpan or @domenic want to give this a final pass? Chrome and Firefox agree on this now.