esid in tests of Stage 3 features

[littledan]

I've seen a bunch of code go in with esid: pending when it's for a feature that hasn't landed yet, and I'm copying this in my PRs for other specs. Is this what we should be doing, or should we expect that the ids on pending specs are stable and just reference them?

[jugglinmike]

Specifying esid: pending serves two purposes: accounting for potential
variability in identifiers prior to formal acceptance, and signaling test
stability to consumers.

Although the first purpose may rarely be necessary, it seems as though it's
always technically possible that IDs change. I'd prefer to play it safe in that
regard.

But if I'm in a minority there, then I would want to make sure we don't drop
the second purpose, as well. In that case, we should introduce a new pending
tag that can be removed when the feature reaches stage 4.

[littledan]

Is there a place we could store metadata relating to the stage of features? Typically when something goes in at Stage 3, it comes with a feature flag; if we could say, "everything with this feature flag is currently at Stage 3", that would seem optimal.

[rwaldron]

@littledan as part of my test262 work over the next 9 months, I will work to get these updated as the features get their esid