Join grapheme clusters on zero-width joiners

1ec5 · Aug 21, 2024 · 5fd311f · 5fd311f
1 parent 6b5c031
commit 5fd311f
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 4 deletions.
diff --git a/build/generate-unicode-data.ts b/build/generate-unicode-data.ts
@@ -36,8 +36,9 @@ fs.writeFileSync('src/data/unicode_properties.ts',
  * Returns whether two grapheme clusters detected by \`Intl.Segmenter\` can be combined to prevent an invisible combining mark from appearing unexpectedly. 
  */
 export function canCombineGraphemes(former: string, latter: string): boolean {
+    // Zero-width joiner
     // Indic_Syllabic_Category=Invisible_Stacker as of Unicode ${indicSyllabicCategory.version}, published ${indicSyllabicCategory.date}.
-    const invisibleStackersRegExp = /[${indicSyllabicCategory.characterClass}]$/u;
-    return invisibleStackersRegExp.test(former) || /^\\p{gc=Mc}/u.test(latter);
+    const terminalJoinersRegExp = /[\u200D${indicSyllabicCategory.characterClass}]$/u;
+    return terminalJoinersRegExp.test(former) || /^\\p{gc=Mc}/u.test(latter);
 }
 `);
diff --git a/src/data/unicode_properties.ts b/src/data/unicode_properties.ts
@@ -4,7 +4,8 @@
  * Returns whether two grapheme clusters detected by `Intl.Segmenter` can be combined to prevent an invisible combining mark from appearing unexpectedly. 
  */
 export function canCombineGraphemes(former: string, latter: string): boolean {
+    // Zero-width joiner
     // Indic_Syllabic_Category=Invisible_Stacker as of Unicode 16.0.0, published 2024-04-30.
-    const invisibleStackersRegExp = /[\u1039\u17D2\u1A60\u1BAB\uAAF6\u{10A3F}\u{11133}\u{113D0}\u{1193E}\u{11A47}\u{11A99}\u{11D45}\u{11D97}\u{11F42}]$/u;
-    return invisibleStackersRegExp.test(former) || /^\p{gc=Mc}/u.test(latter);
+    const terminalJoinersRegExp = /[‍\u1039\u17D2\u1A60\u1BAB\uAAF6\u{10A3F}\u{11133}\u{113D0}\u{1193E}\u{11A47}\u{11A99}\u{11D45}\u{11D97}\u{11F42}]$/u;
+    return terminalJoinersRegExp.test(former) || /^\p{gc=Mc}/u.test(latter);
 }