responsexml-document-properties.htm does not seem supported by specs

[domenic]

https://github.com/w3c/web-platform-tests/blob/master/XMLHttpRequest/responsexml-document-properties.htm

No existing specs state that documents returned from XHRs should have a bunch of undefined properties. What's the deal? @annevk @hallvors

[annevk]

Yeah, it's in fact the opposite.

@hallvors changed the expectations in e3ce86d without consulting with me (I think I'm the original author of this test).

[hallvors]

Hm..I think I asked around for opinions, not sure if I heard back from @annevk or not and I won't spend time looking up E-mails. (I was at Opera back then anyway, don't have that mailbox anymore so it might simply be lost). I think the conclusion back then was along the lines of "assert what a few browers seem to have converged on, we'll get to standardising this later". The history of the test and why it's written like that isn't important, what it should be asserting is. I'd be open to arguments that this is a test suite for XHR and we should refrain from testing stuff that really belongs elsewhere, like the properties defined on a Document or XMLDocument - if you @domenic will argue we should simplify the test and drop all assertations for document properties I can agree that would make some sense. It's still somewhat useful to check what interface/prototype browsers use for xhr.responseXML documents - coding suggestions welcome.

[annevk]

I think we should assert those properties, we should just make them assert the correct values. Which e.g., for document.domain is the empty string.

[annevk]

We should first sort whatwg/dom#221 before we can fix this.