The test for https://github.com/tc39/ecma402/pull/724 (added in
https://github.com/tc39/test262/pull/4328) didn't take the Time Zone
Canonicalization proposal into account; but it should, because that
proposal is stage 3.
As of that proposal, the [[TimeZone]] slot of DateTimeFormat gets the
case-regularized original identifier, no longer the primary identifier. So
the resolvedOptions().timeZone property also no longer returns the primary
identifier.
