fix: French number parsing ambiguity (#7966)

* fix: French number parsing ambiguity

* make code more readable and add script to help verify in future

Author: snowystinger

packages/@internationalized/number/src/NumberParser.ts

@@ -201,10 +201,11 @@ class NumberParserImpl {
       }
     }
 
-    // fr-FR group character is char code 8239, but that's not a key on the french keyboard,
-    // so allow 'period' as a group char and replace it with a space
-    if (this.options.locale === 'fr-FR') {
-      value = replaceAll(value, '.', String.fromCharCode(8239));
+    // fr-FR group character is narrow non-breaking space, char code 8239 (U+202F), but that's not a key on the french keyboard,
+    // so allow space and non-breaking space as a group char as well
+    if (this.options.locale === 'fr-FR' && this.symbols.group) {
+      value = replaceAll(value, ' ', this.symbols.group);
+      value = replaceAll(value, /\u00A0/g, this.symbols.group);
     }
 
     return value;
@@ -303,7 +304,7 @@ function getSymbols(locale: string, formatter: Intl.NumberFormat, intlOptions: I
   return {minusSign, plusSign, decimal, group, literals, numeral, index};
 }
 
-function replaceAll(str: string, find: string, replace: string) {
+function replaceAll(str: string, find: string | RegExp, replace: string) {
   if (str.replaceAll) {
     return str.replaceAll(find, replace);
   }

scripts/checkGroupSeparators.mjs

@@ -0,0 +1,74 @@
+// This script is to run over all of our supported locales and numbering systems
+// and check for the decimal and group separators so we can find
+// non-standard keyboard characters such as the French group separator, narrow non-breaking whitespace.
+// This way we can special case to be more permissive in NumberParser.
+
+
+const NUMBERING_SYSTEMS = ['latn', 'arab', 'hanidec', 'deva', 'beng'];
+let locales = [
+  {label: 'French (France)', value: 'fr-FR'},
+  {label: 'French (Canada)', value: 'fr-CA'},
+  {label: 'German (Germany)', value: 'de-DE'},
+  {label: 'English (Great Britain)', value: 'en-GB'},
+  {label: 'English (United States)', value: 'en-US'},
+  {label: 'Japanese (Japan)', value: 'ja-JP'},
+  {label: 'Danish (Denmark)', value: 'da-DK'},
+  {label: 'Dutch (Netherlands)', value: 'nl-NL'},
+  {label: 'Finnish (Finland)', value: 'fi-FI'},
+  {label: 'Italian (Italy)', value: 'it-IT'},
+  {label: 'Norwegian (Norway)', value: 'nb-NO'},
+  {label: 'Spanish (Spain)', value: 'es-ES'},
+  {label: 'Swedish (Sweden)', value: 'sv-SE'},
+  {label: 'Portuguese (Brazil)', value: 'pt-BR'},
+  {label: 'Chinese (Simplified)', value: 'zh-CN'},
+  {label: 'Chinese (Traditional)', value: 'zh-TW'},
+  {label: 'Korean (Korea)', value: 'ko-KR'},
+  {label: 'Bulgarian (Bulgaria)', value: 'bg-BG'},
+  {label: 'Croatian (Croatia)', value: 'hr-HR'},
+  {label: 'Czech (Czech Republic)', value: 'cs-CZ'},
+  {label: 'Estonian (Estonia)', value: 'et-EE'},
+  {label: 'Hungarian (Hungary)', value: 'hu-HU'},
+  {label: 'Latvian (Latvia)', value: 'lv-LV'},
+  {label: 'Lithuanian (Lithuania)', value: 'lt-LT'},
+  {label: 'Polish (Poland)', value: 'pl-PL'},
+  {label: 'Romanian (Romania)', value: 'ro-RO'},
+  {label: 'Russian (Russia)', value: 'ru-RU'},
+  {label: 'Serbian (Serbia)', value: 'sr-SP'},
+  {label: 'Slovakian (Slovakia)', value: 'sk-SK'},
+  {label: 'Slovenian (Slovenia)', value: 'sl-SI'},
+  {label: 'Turkish (Turkey)', value: 'tr-TR'},
+  {label: 'Ukrainian (Ukraine)', value: 'uk-UA'},
+  {label: 'Arabic (United Arab Emirates)', value: 'ar-AE'}, // ar-SA??
+  {label: 'Greek (Greece)', value: 'el-GR'},
+  {label: 'Hebrew (Israel)', value: 'he-IL'}
+];
+
+let separators = new Map();
+for (let nums of NUMBERING_SYSTEMS) {
+  for (let locale of locales) {
+    let formatter = new Intl.NumberFormat(locale.value + '-u-nu-' + nums, {});
+    let parts = formatter.formatToParts(10000000000);
+    let separator = parts.find(p => p.type === 'group')?.value;
+    if (separators.has(separator)) {
+      separators.set(separator, [...separators.get(separator), locale.value + '-u-nu-' + nums])
+    } else {
+      separators.set(separator, [separator.charCodeAt(0), locale.value + '-u-nu-' + nums]);
+    }
+  }
+}
+console.log(separators);
+
+let decimals = new Map();
+for (let nums of NUMBERING_SYSTEMS) {
+  for (let locale of locales) {
+    let formatter = new Intl.NumberFormat(locale.value + '-u-nu-' + nums, {minimumFractionDigits: 2});
+    let parts = formatter.formatToParts(10000.0000001);
+    let decimal = parts.find(p => p.type === 'decimal')?.value;
+    if (decimals.has(decimal)) {
+      decimals.set(decimal, [...decimals.get(decimal), locale.value + '-u-nu-' + nums])
+    } else {
+      decimals.set(decimal, [decimal.charCodeAt(0), locale.value + '-u-nu-' + nums]);
+    }
+  }
+}
+console.log(decimals);