Allow Unicode minus sign (U+2212) in ISO 8601 strings

Wherever a sign character is needed (in 6-digit extended years and in
time zone offsets), also accept the Unicode minus sign as well as the
normal ASCII + and -.

Closes: #814

polyfill/lib/ecmascript.mjs

@@ -96,7 +96,9 @@ export const ES = ObjectAssign({}, ES2019, {
     const regex = zoneRequired ? PARSE.absolute : PARSE.datetime;
     const match = regex.exec(isoString);
     if (!match) throw new RangeError(`invalid ISO 8601 string: ${isoString}`);
-    const year = ES.ToInteger(match[1]);
+    let yearString = match[1];
+    if (yearString[0] === '\u2212') yearString = `-${yearString.slice(1)}`;
+    const year = ES.ToInteger(yearString);
     const month = ES.ToInteger(match[2] || match[4]);
     const day = ES.ToInteger(match[3] || match[5]);
     const hour = ES.ToInteger(match[6]);
@@ -107,8 +109,9 @@ export const ES = ObjectAssign({}, ES2019, {
     const millisecond = ES.ToInteger(fraction.slice(0, 3));
     const microsecond = ES.ToInteger(fraction.slice(3, 6));
     const nanosecond = ES.ToInteger(fraction.slice(6, 9));
-    const offset = `${match[14]}:${match[15] || '00'}`;
-    let ianaName = match[16];
+    const offsetSign = match[14] === '-' || match[14] === '\u2212' ? '-' : '+';
+    const offset = `${offsetSign}${match[15]}:${match[16] || '00'}`;
+    let ianaName = match[17];
     if (ianaName) {
       try {
         // Canonicalize name if it is an IANA link name or is capitalized wrong
@@ -118,7 +121,7 @@ export const ES = ObjectAssign({}, ES2019, {
       }
     }
     const zone = match[13] ? 'UTC' : ianaName || offset;
-    const calendar = match[17] || null;
+    const calendar = match[18] || null;
     return {
       year,
       month,
@@ -167,7 +170,9 @@ export const ES = ObjectAssign({}, ES2019, {
     const match = PARSE.yearmonth.exec(isoString);
     let year, month, calendar, refISODay;
     if (match) {
-      year = ES.ToInteger(match[1]);
+      let yearString = match[1];
+      if (yearString[0] === '\u2212') yearString = `-${yearString.slice(1)}`;
+      year = ES.ToInteger(yearString);
       month = ES.ToInteger(match[2]);
       calendar = match[3] || null;
     } else {
@@ -1572,7 +1577,7 @@ const OFFSET = new RegExp(`^${PARSE.offset.source}$`);
 function parseOffsetString(string) {
   const match = OFFSET.exec(String(string));
   if (!match) return null;
-  const sign = match[1] === '-' ? -1 : +1;
+  const sign = match[1] === '-' || match[1] === '\u2212' ? -1 : +1;
   const hours = +match[2];
   const minutes = +(match[3] || 0);
   return sign * (hours * 60 + minutes) * 60 * 1e9;

polyfill/lib/regex.mjs

@@ -1,7 +1,8 @@
-const yearpart = /(?:[+-]\d{6}|\d{4})/;
+const yearpart = /(?:[+-\u2212]\d{6}|\d{4})/;
 const datesplit = new RegExp(`(${yearpart.source})(?:-(\\d{2})-(\\d{2})|(\\d{2})(\\d{2}))`);
 const timesplit = /(\d{2})(?::(\d{2})(?::(\d{2})(?:[.,](\d{1,9}))?)?|(\d{2})(?:(\d{2})(?:[.,](\d{1,9}))?)?)?/;
-const zonesplit = /(?:([zZ])|(?:([+-]\d{2})(?::?(\d{2}))?(?:\[(?!c=)([^\]\s]*)?\])?))/;
+export const offset = /([+-\u2212])([0-2][0-9])(?::?([0-5][0-9]))?/;
+const zonesplit = new RegExp(`(?:([zZ])|(?:${offset.source}?(?:\\[(?!c=)([^\\]\\s]*)?\\])?))`);
 const calendar = /\[c=([^\]\s]+)\]/;
 
 export const absolute = new RegExp(
@@ -23,5 +24,4 @@ export const time = new RegExp(`^${timesplit.source}(?:${zonesplit.source})?(?:$
 export const yearmonth = new RegExp(`^(${yearpart.source})-?(\\d{2})$`);
 export const monthday = /^(?:--)?(\d{2})-?(\d{2})$/;
 
-export const offset = /([+-])([0-2][0-9])(?::?([0-5][0-9]))?/;
 export const duration = /^P(?:(\d+)Y)?(?:(\d+)M)?(?:(\d+)W)?(?:(\d+)D)?(?:T(?!$)(?:(\d+)H)?(?:(\d+)M)?(?:(\d+)(?:[.,](\d{1,9}))?S)?)?$/i;

polyfill/lib/timezone.mjs

@@ -26,7 +26,7 @@ const OFFSET = new RegExp(`^${REGEX.offset.source}$`);
 function parseOffsetString(string) {
   const match = OFFSET.exec(String(string));
   if (!match) return null;
-  const sign = match[1] === '-' ? -1 : +1;
+  const sign = match[1] === '-' || match[1] === '\u2212' ? -1 : +1;
   const hours = +match[2];
   const minutes = +(match[3] || 0);
   return sign * (hours * 60 + minutes) * 60 * 1e9;

polyfill/test/absolute.mjs

@@ -323,6 +323,10 @@ describe('Absolute', () => {
     it('variant decimal separator', () => {
       equal(`${Absolute.from('1976-11-18T15:23:30,12Z')}`, '1976-11-18T15:23:30.120Z');
     });
+    it('variant minus sign', () => {
+      equal(`${Absolute.from('1976-11-18T15:23:30.12\u221202:00')}`, '1976-11-18T17:23:30.120Z');
+      equal(`${Absolute.from('\u2212009999-11-18T15:23:30.12Z')}`, '-009999-11-18T15:23:30.120Z');
+    });
     it('mixture of basic and extended format', () => {
       equal(`${Absolute.from('19761118T15:23:30.1+00:00')}`, '1976-11-18T15:23:30.100Z');
       equal(`${Absolute.from('1976-11-18T152330.1+00:00')}`, '1976-11-18T15:23:30.100Z');

polyfill/test/datetime.mjs

@@ -473,6 +473,10 @@ describe('DateTime', () => {
     it('variant decimal separator', () => {
       equal(`${DateTime.from('1976-11-18T15:23:30,12Z')}`, '1976-11-18T15:23:30.120');
     });
+    it('variant minus sign', () => {
+      equal(`${DateTime.from('1976-11-18T15:23:30.12\u221202:00')}`, '1976-11-18T15:23:30.120');
+      equal(`${DateTime.from('\u2212009999-11-18T15:23:30.12')}`, '-009999-11-18T15:23:30.120');
+    });
     it('mixture of basic and extended format', () => {
       equal(`${DateTime.from('1976-11-18T152330.1+00:00')}`, '1976-11-18T15:23:30.100');
       equal(`${DateTime.from('19761118T15:23:30.1+00:00')}`, '1976-11-18T15:23:30.100');

polyfill/test/regex.mjs

@@ -57,6 +57,11 @@ describe('fromString regex', () => {
     generateTest('1976-11-18T15:23', 'z', [1976, 11, 18, 15, 23, 30, 123, 456, 789]);
     // Comma decimal separator
     test('1976-11-18T15:23:30,1234Z', [1976, 11, 18, 15, 23, 30, 123, 400]);
+    // Unicode minus sign
+    ['\u221204:00', '\u221204', '\u22120400'].forEach((offset) =>
+      test(`1976-11-18T15:23:30.1234${offset}`, [1976, 11, 18, 19, 23, 30, 123, 400])
+    );
+    test('\u2212009999-11-18T15:23:30.1234Z', [-9999, 11, 18, 15, 23, 30, 123, 400]);
     // Mixture of basic and extended format
     test('1976-11-18T152330Z', [1976, 11, 18, 15, 23, 30]);
     test('1976-11-18T152330.1234Z', [1976, 11, 18, 15, 23, 30, 123, 400]);
@@ -112,6 +117,11 @@ describe('fromString regex', () => {
     test('1976-11-18T15:23:30.12345678', [1976, 11, 18, 15, 23, 30, 123, 456, 780]);
     // Comma decimal separator
     test('1976-11-18T15:23:30,1234', [1976, 11, 18, 15, 23, 30, 123, 400]);
+    // Unicode minus sign
+    ['\u221204:00', '\u221204', '\u22120400'].forEach((offset) =>
+      test(`1976-11-18T15:23:30.1234${offset}`, [1976, 11, 18, 15, 23, 30, 123, 400])
+    );
+    test('\u2212009999-11-18T15:23:30.1234', [-9999, 11, 18, 15, 23, 30, 123, 400]);
     // Mixture of basic and extended format
     test('1976-11-18T152330', [1976, 11, 18, 15, 23, 30]);
     test('1976-11-18T152330.1234', [1976, 11, 18, 15, 23, 30, 123, 400]);
@@ -164,6 +174,8 @@ describe('fromString regex', () => {
       '19761118T152330',
       '19761118T152330.1234'
     ].forEach((str) => test(str, [1976, 11, 18]));
+    // Unicode minus sign
+    test('\u2212009999-11-18', [-9999, 11, 18]);
     // Representations with reduced precision
     test('1976-11-18T15', [1976, 11, 18]);
     // Date-only forms
@@ -277,6 +289,8 @@ describe('fromString regex', () => {
       // Representations with reduced precision
       '1976-11-18T15'
     ].forEach((str) => test(str, [1976, 11]));
+    // Unicode minus sign
+    test('\u2212009999-11-18T15:23:30.1234Z', [-9999, 11]);
     // Date-only forms
     test('1976-11-18', [1976, 11]);
     test('19761118', [1976, 11]);
@@ -331,6 +345,8 @@ describe('fromString regex', () => {
     [
       // Comma decimal separator
       '1976-11-18T15:23:30,1234',
+      // Unicode minus sign
+      '\u2212009999-11-18',
       // Mixture of basic and extended format
       '1976-11-18T152330',
       '1976-11-18T152330.1234',
@@ -396,6 +412,8 @@ describe('fromString regex', () => {
     generateTest('1976-11-18T15:23', 'z', 'UTC');
     // Comma decimal separator
     test('1976-11-18T15:23:30,1234Z', 'UTC');
+    // Unicode minus sign
+    ['\u221204:00', '\u221204', '\u22120400'].forEach((offset) => test(`1976-11-18T15:23${offset}`, '-04:00'));
     [
       // Mixture of basic and extended format
       '1976-11-18T152330',
@@ -425,6 +443,12 @@ describe('fromString regex', () => {
     test('-03:00', '-03:00');
     test('+03', '+03:00');
     test('-03', '-03:00');
+    test('\u22120000', '+00:00');
+    test('\u221200:00', '+00:00');
+    test('\u221200', '+00:00');
+    test('\u22120300', '-03:00');
+    test('\u221203:00', '-03:00');
+    test('\u221203', '-03:00');
     // Representations with calendar
     test('1976-11-18T15:23:30.123456789Z[c=iso8601]', 'UTC');
     test('1976-11-18T15:23:30.123456789-04:00[c=iso8601]', '-04:00');

polyfill/test/time.mjs

@@ -393,6 +393,9 @@ describe('Time', () => {
       it('variant decimal separator', () => {
         equal(`${Time.from('1976-11-18T15:23:30,12Z')}`, '15:23:30.120');
       });
+      it('variant minus sign', () => {
+        equal(`${Time.from('1976-11-18T15:23:30.12\u221202:00')}`, '15:23:30.120');
+      });
       it('basic format', () => {
         equal(`${Time.from('152330')}`, '15:23:30');
         equal(`${Time.from('152330.1')}`, '15:23:30.100');

polyfill/test/timezone.mjs

@@ -47,6 +47,9 @@ describe('TimeZone', () => {
     test('+0330');
     test('-0650');
     test('-08');
+    test('\u221201:00');
+    test('\u22120650');
+    test('\u221208');
     test('Europe/Vienna');
     test('America/New_York');
     test('Africa/CAIRO'); // capitalization
@@ -63,6 +66,9 @@ describe('TimeZone', () => {
     test('+0330', '+03:30');
     test('-0650', '-06:50');
     test('-08', '-08:00');
+    test('\u221201:00', '-01:00');
+    test('\u22120650', '-06:50');
+    test('\u221208', '-08:00');
     test('Europe/Vienna');
     test('America/New_York');
     test('Africa/CAIRO', 'Africa/Cairo');
@@ -79,6 +85,9 @@ describe('TimeZone', () => {
     test('+0330');
     test('-0650');
     test('-08');
+    test('\u221201:00');
+    test('\u22120650');
+    test('\u221208');
     test('Europe/Vienna');
     test('America/New_York');
     test('Africa/CAIRO');
@@ -102,6 +111,9 @@ describe('TimeZone', () => {
     test('1994-11-05T08:15:30-05:00', '-05:00');
     test('1994-11-05T08:15:30-05:00[America/New_York]', 'America/New_York');
     test('1994-11-05T08:15:30-05[America/New_York]', 'America/New_York');
+    test('1994-11-05T08:15:30\u221205:00', '-05:00');
+    test('1994-11-05T08:15:30\u221205:00[America/New_York]', 'America/New_York');
+    test('1994-11-05T08:15:30\u221205[America/New_York]', 'America/New_York');
     test('1994-11-05T13:15:30Z', 'UTC');
     function test(isoString, name) {
       const tz = Temporal.TimeZone.from(isoString);

polyfill/test/validStrings.mjs

@@ -152,7 +152,7 @@ function seq(...productions) {
 // Grammar productions, based on the grammar in RFC 3339
 
 // characters
-const sign = character('+-');
+const sign = character('+-−');
 const decimalSeparator = character('.,');
 const daysDesignator = character('Dd');
 const hoursDesignator = character('Hh');
@@ -172,7 +172,10 @@ const utcDesignator = withCode(character('Zz'), (data) => {
 
 const dateFourDigitYear = repeat(4, digit());
 const dateExtendedYear = seq(sign, repeat(6, digit()));
-const dateYear = withCode(choice(dateFourDigitYear, dateExtendedYear), (data, result) => (data.year = +result));
+const dateYear = withCode(
+  choice(dateFourDigitYear, dateExtendedYear),
+  (data, result) => (data.year = +result.replace('\u2212', '-'))
+);
 const dateMonth = withCode(zeroPaddedInclusive(1, 12, 2), (data, result) => (data.month = +result));
 const dateDay = withCode(zeroPaddedInclusive(1, 31, 2), (data, result) => (data.day = +result));
 
@@ -189,7 +192,10 @@ const timeFractionalPart = withCode(between(1, 9, digit()), (data, result) => {
   data.nanosecond = +fraction.slice(6, 9);
 });
 const timeFraction = seq(decimalSeparator, timeFractionalPart);
-const timeZoneUTCOffsetSign = withCode(sign, (data, result) => (data.offsetSign = result));
+const timeZoneUTCOffsetSign = withCode(
+  sign,
+  (data, result) => (data.offsetSign = result === '-' || result === '\u2212' ? '-' : '+')
+);
 const timeZoneUTCOffsetHour = withCode(zeroPaddedInclusive(0, 23, 2), (data, result) => (data.offsetHour = +result));
 const timeZoneUTCOffsetMinute = withCode(
   zeroPaddedInclusive(0, 59, 2),

polyfill/test/yearmonth.mjs

@@ -75,6 +75,10 @@ describe('YearMonth', () => {
         equal(`${YearMonth.from('197611')}`, '1976-11');
         equal(`${YearMonth.from('+00197611')}`, '1976-11');
       });
+      it('variant minus sign', () => {
+        equal(`${YearMonth.from('\u2212009999-11')}`, '-009999-11');
+        equal(`${YearMonth.from('1976-11-18T15:23:30.1\u221202:00')}`, '1976-11');
+      });
       it('mixture of basic and extended format', () => {
         equal(`${YearMonth.from('1976-11-18T152330.1+00:00')}`, '1976-11');
         equal(`${YearMonth.from('19761118T15:23:30.1+00:00')}`, '1976-11');

spec/abstractops.html

@@ -214,8 +214,10 @@ <h1>ISO 8601 grammar</h1>
       NonzeroDigit : one of
           `1` `2` `3` `4` `5` `6` `7` `8` `9`
 
-      Sign : one of
-          `+` `-`
+      Sign :
+          `+`
+          `-`
+          U+2212
 
       DecimalSeparator : one of
           `.` `,`
@@ -447,6 +449,7 @@ <h1>ParseISODateTime ( _isoString_ )</h1>
       1. Assert: Type(_isoString_) is String.
       1. Let _year_, _month_, _day_, _hour_, _minute_, _second_, and _fraction_ be the parts of _isoString_ produced respectively by the |DateYear|, |DateMonth|, |DateDay|, |TimeHour|, |TimeMinute|, |TimeSecond|, and |TimeFractionalPart| productions, or *undefined* if not present.
       1. Let _year_ be the part of _isoString_ produced by the |DateYear| production.
+      1. If the first code unit of _year_ is 0x2212 (MINUS SIGN), replace it with the code unit 0x002D (HYPHEN-MINUS).
       1. Set _year_ to ! ToInteger(_year_).
       1. If _month_ is *undefined*, then
         1. Set _month_ to 1.
@@ -667,7 +670,7 @@ <h1>ParseTemporalTimeZoneString ( _isoString_ )</h1>
       1. If _hour_ is not *undefined*, then
         1. Assert: _sign_ is not *undefined*.
         1. Set _hour_ to ! ToInteger(_hour_).
-        1. If _sign_ = `"-"`, then
+        1. If _sign_ is the code unit 0x002D (HYPHEN-MINUS) or the code unit 0x2212 (MINUS SIGN), then
           1. Set _sign_ to −1.
         1. Else,
           1. Set _sign_ to 1.