Editorial: Update GetOption to throw a TypeError when a required option is missing

[gibson042]

TypeError rather than RangeError is in better alignment with other parts of the spec where a required parameter or option is missing:

• ECMA-262 GroupBy
• ECMA-262 Object static functions
• ECMA-262 RegExp.escape
• ECMA-262 Array.prototype.every et al.
  • Also {Map,Set}.prototype.{forEach,getOrInsertComputed}
• ECMA-262 Uint8Array static methods
• ECMA-262 Set.prototype.{difference,intersection,union,...}
• ECMA-262 WeakMap methods and WeakSet methods
• ECMA-262 Promise
• ECMA-262 Reflect functions
• Intl.DateTimeFormat.prototype.{formatRange,formatRangeToParts}
  • Also Intl.NumberFormat.prototype.{formatRange,formatRangeToParts} and Intl.PluralRules.prototype.selectRange
• Intl.DisplayNames
• Intl.DurationFormat.prototype.{format,formatToParts}
• Intl.Locale
• SetNumberFormatUnitOptions as exercised by e.g. new Intl.NumberFormat("en", { style: "currency" }) and new Intl.NumberFormat("en", { style: "unit" })

Fortunately, this change is editorial because no current site calls GetOption with default ~required~, although I expect proposals like Amount to do so.

[anba]

This is a normative change for Temporal:

// Currently spec'ed to throw a RangeError.
new Temporal.ZonedDateTime(0n, "UTC").getTimeZoneTransition({})

The required option was added to GetOption in #687.

[gibson042]

The required option was added to GetOption in #687.

Yes, I remember that... most of #687 was subsequently reverted in #822, but not that part.

This is a normative change for Temporal:

// Currently spec'ed to throw a RangeError.
new Temporal.ZonedDateTime(0n, "UTC").getTimeZoneTransition({})

AFAICT, it's actually not... I don't see any uses of GetOption in #1044, and tc39/ecma262#3759 is no longer introducing GetOption to ECMA-262, instead having its own GetTemporalUnitValuedOption for reading unit/smallestUnit/largestUnit properties and GetDirectionOption for handling direction in Temporal.ZonedDateTime.prototype.getTimeZoneTransition (the latter being responsible for the above RangeError).

But you're right to single out Temporal here, and it is odd that it diverges in this way. It seems like the Temporal RangeErrors were introduced without scrutiny in ToSmallestTemporalUnit by tc39/proposal-temporal#886 (where they were appropriate for «If disallowedUnits contains smallestUnit» but not for «If smallestUnit is "not present"») and Temporal.Duration.prototype.total by tc39/proposal-temporal#1064 (later moved into ToTemporalDurationTotalUnit by tc39/proposal-temporal#1599), carried forward without comment into GetTemporalUnit by tc39/proposal-temporal#2248 , rolled into GetOption by #687 , and then Temporal subsequently abandoned GetOption but kept the behavior.

I'm still very much in favor of having the more standard behavior here, and will see what can be done to bring Temporal in line (although it looks like there is some test262 coverage:

• built-ins/Temporal/ZonedDateTime/prototype/getTimeZoneTransition/direction-undefined.js
• built-ins/Temporal/Instant/prototype/round/throws-without-smallest-unit.js
• built-ins/Temporal/PlainDateTime/prototype/round/throws-argument-object.js
• built-ins/Temporal/PlainTime/prototype/round/smallestunit-missing.js
• built-ins/Temporal/ZonedDateTime/prototype/round/throws-without-smallestunit.js

[anba]

AFAICT, it's actually not... I don't see any uses of GetOption in #1044, [...]

Ah, I see. I didn't know that this changed in the ECMA262 spec PR when compared to what's on https://tc39.es/proposal-temporal.

[gibson042]

Ah, I see. I didn't know that this changed in the ECMA262 spec PR when compared to what's on https://tc39.es/proposal-temporal.

Yeah, we're in an awkward state right now where the source of truth for Temporal is at tc39/ecma262#3759 (and its ECMA-402 sibling at #1044 ).

[sffc]

Yes, if a required option is missing, it is a TypeError. If Temporal is doing something different, then it is a bug in Temporal.

[sffc]

TG2 agreement: https://github.com/tc39/ecma402/blob/main/meetings/notes-2026-06-11.md#editorial-revise-getoption-to-trigger-a-typeerror-for-absent-mandatory-options-potentially-significant-enough-for-collective-review