[WIP] Update testing of Request's default credentials mode

[domfarolino]

This is for whatwg/fetch#585.

I'm a little surprised if only these tests have to be updated, so I'm marking this as a WIP for now. Basically what I did though, was:

• Ran all of the /fetch, /xor, service-workers, and /cors tests on Chromium, and noted failues (none)
• Modified the Request constructor to use the same-origin credentials fallback instead of the previous omit
• Ran all of the previous tests again and noted the failures (only two)
• Updated the affected tests!

[annevk]

I guess we don't really have good same-origin cookie/HTTP authentication tests then.

@wanderview does that make sense?

[ghost]

Build PASSED

Started: 2018-03-08 11:12:38
Finished: 2018-03-08 11:23:12

Failing Jobs

• MicrosoftEdge:16.16299

View more information about this build on:

• WPT PR Status
• TravisCI

[domfarolino]

Pinging @wanderview here. This seems to be the only thing blocking Chrome's implementation so the sooner we can get this in the better :)

[wanderview]

LGTM. I'm surprised there aren't any service worker tests that check the default credentials via the FetchEvent.

[domfarolino]

Thanks a lot! Me too, should we make an issue to add some at some point? I would expect more tests to be failing here, and we probably want more tests to be doing so.

[annevk]

Thoughts on whatwg/fetch#585 (comment) anyone? Sounds like @wanderview wants to go ahead with this in Firefox as well, so I'm happy to land this, just wondering about follow-up work and how to structure that.

[wanderview]

BTW, the service worker tests mostly check that the correct value is passed through to FetchEvent. The tests use explicit modes for this. They are not really checking for defaults. I think that its reasonable for the fetch tests to mainly be checking the defaults.