Normative: Remove implementation-defined typeof behavior

[littledan]

Previously, the ECMAScript specification permitted non-callable
non-standard exotic objects to have implementation-defined typeof,
within certain limits. Although some ECMAScript implementations
may have taken advantage of this possibility historically, it's not
clear that anyone needs it anymore. Instead, other mechanisms can
be used by embedders to indicate aspects of objects.

Closes #1440

[ljharb]

Pretty hard to test a lack of unknown implementation-defined behavior here, so I don't think this would need any tests :-)

[littledan]

I'm wondering about the "needs consensus" label here. Do you think this needs to be presented at a TC39 meeting, or could we get consensus for it on GitHub?

[ljharb]

Generally, every normative change requires committee approval; it certainly seems like something we could decide on GitHub, but I would be uncomfortable merging a non-bugfix normative change without that explicit consensus.

[allenwb]

Instead, @@toStringTag and other
mechanisms can be used by embedders to indicate aspects of objects

Just wanted to remind that @@toStringTag, unlike typeof, is forgeable and was never intended to be used as a reliable type checking mechanism. We shouldn't encourage people to try to use it that way.

Otherwise, I like the change this issue is proposing.

[annevk]

You should also remove the note below the table as it's no longer applicable.

[littledan]

Removed the reference to @@toStringTag as @allenwb suggested and the note that @annevk pointed out. Thanks for the quick reviews everyone!

To the editor group: Do you plan to put this on the agenda for the upcoming TC39 meeting, or should I do it?

[ljharb]

It’s always preferred when committee members can drive their own needs-consensus changes.

[littledan]

Alright, thanks for the clarification and quick review.

[jmdyck]

It’s always preferred when committee members can drive their own needs-consensus changes.

So what should happen when a needs-consensus change doesn't 'belong' to a committee member?

[ljharb]

@jmdyck in that case (or when no committee member is driving it) it's up to the editors to proactively add it to the agenda, as time permits.

[jmdyck]

Cool, thanks.

[ljharb]

The HTML change for this seems to have landed whatwg/webidl#640

[annevk]

To be clear, that text can be removed if this lands. It's mostly added there since this change is taking a while.

note has been removed