Report EventSource & track ResourceTiming entries

[noamr]

• At least two implementers are interested (and none opposed):

  • Already implemented

• Tests are written and can be reviewed and commented upon at:

  • EventSource RT test
  • Track initiator type test - tests the existence and validity of RT entries for <track>.

• Implementation bugs are filed:

  • N/A

(See WHATWG Working Mode: Changes for more details.)

---

/infrastructure.html ( diff )
/media.html ( diff )
/server-sent-events.html ( diff )

[domenic]

Oof. I am not too comfortable with this meshing of the rigorous processResponseEndOfBody with the non-rigorous paragraphs below this, even if you have a note paragraph explaining the mismatch. Honestly I'm not too happy with the rigorous request setup and fetch call, when in reality it sounds like there will be multiple fetch calls all of which share some characteristics of the request?

I don't really know how to square this. I'd like to call in @annevk for help.

[noamr]

Yea I was surprised that the whole range-request thing is a free-for-all which makes it hard to add reporting on top.

[noamr]

Perhaps we should state here that user-agents may make several rigorous fetch calls that look a certain way with the optional range headers.

[annevk]

Yeah, that's the desired setup (though with certain limitations). @jakearchibald worked on this in #2814 but wasn't able to finish it. whatwg/fetch#144 (comment) also describes some of the complexity here.

The basic setup is that there needs to be an initial request (that gets the start of the response to scan) and identifies the post-redirect response URL and then zero or more subsequent requests (that go directly to the post-redirect response URL from the initial request). Perhaps writing that down and iterating on it would be a good way to go here?

@padenot has been looking at this in Firefox to address some security issues.

[noamr]

I don't mind taking this on. Since this is becoming an issue of its own though, I'm thinking to split it it out from the EventSource patch.

[domenic]

Big <3 for doing the extra work of making EventSource rigorous, instead of just taking my suggestion to move all the non-rigorous stuff into the constructor.

[domenic]

Are there tests for <track>? We can merge this after the template in the OP is filled out.