typo (no timeout after load)

[Mouvedia]

<TimothyGu> Could you file an issue?

972797f#r29848015

[annevk]

I think completed is fine here, since we don't know if the request will succeed or fail.

[annevk]

@TimothyGu I guess you had a different perspective on that?

[Mouvedia]

You are using the term "completed" for onloaded which means abort can fire after onload; surely that's not desired? The order matters.

cf https://xhr.spec.whatwg.org/#ref-for-concept-event-fire-progress%E2%91%A4

[annevk]

This table has no normative meaning, it just summarizes a bunch of details. "Completed" just means that the fetch is over with.

[TimothyGu]

@annevk Nah I wasn't sure if this was an actual bug or no so I just asked @Mouvedia to file an issue.

[Mouvedia]

@annevk it's unclear and was incorrectly implemented before: https://bugzilla.mozilla.org/show_bug.cgi?id=703380

[annevk]

@Mouvedia that bug doesn't cite the specification as being problematic though. And again, it's a non-normative summary that implementers shouldn't really consult anyway.

I suppose we could add a paragraph after the table: "The abort, error, load, and timeout events are mutually exclusive." Would that help?

[Mouvedia]

I suppose we could add a paragraph after the table: "The abort, error, load, and timeout events are mutually exclusive." Would that help?

That would be even more confusing. Forget about it.