Add coverage for uvalue "true" and "yes"

[anba]

Ensure yes is only canonicalized to true for some ukeys.

[anba]

"constructor-unicode-extension-uvalue-yes.js" fails in JSC and V8 because of ICU4C bugs.

[anba]

No other test file in "test/intl402" mentions either for-of or generators, so for consistency let's omit them here, too.

[github-actions]

2 new or modified tests were run on 7 engines.

View workflow run

Test	Engine262	GraalJS	JavaScriptCore	QuickJS	SpiderMonkey	V8	XS
intl402/Locale/constructor-unicode-extension-uvalue-true.js	❌	✅	✅	❌	✅	✅	❌
intl402/Locale/constructor-unicode-extension-uvalue-yes.js	❌	✅	❌	❌	✅	❌	❌

[jugglinmike]

No conflicts and no overlap with more recent additions; this all looks good to me!

For those curious: here's the source for which fields have "yes" canonicalized to "true".

(We're missing a bit of coverage, but this patch has waited long enough--I'll follow up with one of my own.)

[gibson042]

ukey can start with a digit.

Suggested change

	// Generate all possible `ukey` values.
	function* ukeys() {
	const lowerA = 'a'.charCodeAt(0);
	const lowerZ = 'z'.charCodeAt(0);
	for (let first = lowerA; first <= lowerZ; ++first) {
	for (let second = lowerA; second <= lowerZ; ++second) {
	yield String.fromCharCode(first, second);
	}
	}
	}
	// Generate all possible `ukey` values.
	// https://unicode.org/reports/tr35/#ukey
	function* ukeys() {
	const lowerAlpha = 'abcdefghijklmnopqrstuvwxyz';
	const lowerAlphanum = lowerAlpha + '0123456789';
	for (let i = 0; i < lowerAlphanum.length; ++i) {
	for (let j = 0; j < lowerAlpha.length; ++j) {
	yield lowerAlphanum[i] + lowerAlpha[j];
	}
	}
	}

[gibson042]

Suggestion:

Suggested change

	assert.sameValue(
	new Intl.Locale(`en-u-${ukey}-true`).toString(),
	`en-u-${ukey}`,
	`new Intl.Locale("en-u-${ukey}-true").toString() returns "en-u-${ukey}"`
	);
	const localeId = `en-u-${ukey}-true`;
	const normalizedLocaleId = `en-u-${ukey}`;
	assert.sameValue(
	new Intl.Locale(localeId).toString(),
	normalizedLocaleId,
	`new Intl.Locale("${localeId}").toString() must be "${normalizedLocaleId}"`
	);

[gibson042]

Suggestion:

Suggested change

	function canonicalizeUValue(ukey, uvalue) {
	assert.sameValue(uvalue, "yes", "unexpected uvalue");
	switch (ukey) {
	case "kb":
	case "kc":
	case "kh":
	case "kk":
	case "kn":
	// Canonicalized to "true", which then gets removed.
	return "";
	default:
	return uvalue;
	}
	}
	for (let ukey of ukeys()) {
	let canonicalized = new Intl.Locale(`en-u-${ukey}-yes`).toString();
	if (canonicalizeUValue(ukey, "yes") === "yes") {
	assert.sameValue(
	canonicalized,
	`en-u-${ukey}-yes`,
	`new Intl.Locale("en-u-${ukey}-yes").toString() returns "en-u-${ukey}-yes"`
	);
	} else {
	assert.sameValue(
	canonicalized,
	`en-u-${ukey}`,
	`new Intl.Locale("en-u-${ukey}-yes").toString() returns "en-u-${ukey}"`
	);
	}
	}
	// Whether ukey has a "yes" alias for "true" in
	// https://github.com/unicode-org/cldr/blob/maint/maint-48/common/bcp47/collation.xml
	function hasAlias(ukey) {
	return ukey === "kb" || ukey === "kc" || ukey === "kh" || ukey === "kk" || ukey === "kn";
	}
	for (let ukey of ukeys()) {
	const localeId = `en-u-${ukey}-yes`;
	const normalizedLocaleId = hasAlias(ukey) ? `en-u-${ukey}` : localeId;
	assert.sameValue(
	new Intl.Locale(localeId).toString(),
	normalizedLocaleId,
	`new Intl.Locale("${localeId}").toString() must be "${normalizedLocaleId}"`
	);
	}

[gibson042]

Suggested change

	// Generate all possible `ukey` values.
	function* ukeys() {
	const lowerA = 'a'.charCodeAt(0);
	const lowerZ = 'z'.charCodeAt(0);
	for (let first = lowerA; first <= lowerZ; ++first) {
	for (let second = lowerA; second <= lowerZ; ++second) {
	yield String.fromCharCode(first, second);
	}
	}
	}
	// Generate all possible `ukey` values.
	// https://unicode.org/reports/tr35/#ukey
	function* ukeys() {
	const lowerAlpha = 'abcdefghijklmnopqrstuvwxyz';
	const lowerAlphanum = lowerAlpha + '0123456789';
	for (let i = 0; i < lowerAlphanum.length; ++i) {
	for (let j = 0; j < lowerAlpha.length; ++j) {
	yield lowerAlphanum[i] + lowerAlpha[j];
	}
	}
	}

[gibson042]

Suggested change

	description: >
	Remove any uvalue (aka type) equal to "true".
	description: >
	Remove any uvalue (aka type) equal to "yes" if and only if "yes" is an alias for "true".

[jugglinmike]

@gibson042 A few details make this a good candidate for landing as-is: your suggestions are additive rather than corrective, this has been waiting for some time, @anba's public contribution history suggests they may not be available right now, and I have immediate availability to make the requested changes. Would you be open to landing this patch as-is with the understanding that I will submit a follow-up later today?

[gibson042]

@jugglinmike Yes, I left a comment rather than a request for changes specifically to avoid blocking.

[jugglinmike]

Right on!

Done

[jugglinmike]

Follow-up patch: #5138