Ensure reference data is emitted when calendarName = 'always'

[jessealama]

Needed change to the polyfill will happen in a separate PR.

References:

• YYYY-MM[u-ca=iso8601] and MM-DD[u-ca=iso8601] formats are not round-trippable proposal-temporal#1971

[jessealama]

cc @Ms2ger @ptomato

[jessealama]

Polyfill change here.

[jessealama]

(Unlike the proposal-temporal repo, I can't request a reviewer when making a PR.)

[Ms2ger]

lgtm once we get consensus

[ptomato]

I think this may also require some updates to existing test262 tests.

[ptomato]

For the record, these tests are related to tc39/proposal-temporal#2035 which is a normative change to the Temporal proposal that I plan to present for consensus at the March TC39 meeting.

I think this may also require some updates to existing test262 tests.

What I meant by this - I believe the following existing tests (and possibly others?) might start to fail if implementations start following that normative change, so I think they should be adapted before this PR can be considered complete.

• built-ins/Temporal/PlainMonthDay/prototype/toString/calendarname-always.js
• built-ins/Temporal/PlainYearMonth/prototype/toString/calendarname-always.js

[jessealama]

For the record, these tests are related to tc39/proposal-temporal#2035 which is a normative change to the Temporal proposal that I plan to present for consensus at the March TC39 meeting.

I think this may also require some updates to existing test262 tests.

What I meant by this - I believe the following existing tests (and possibly others?) might start to fail if implementations start following that normative change, so I think they should be adapted before this PR can be considered complete.

* built-ins/Temporal/PlainMonthDay/prototype/toString/calendarname-always.js

* built-ins/Temporal/PlainYearMonth/prototype/toString/calendarname-always.js

Good catch -- I've updated both of these.

[ptomato]

The corresponding normative change achieved consensus, so I'll merge this.