fix Extend + Extended_Pictographic sequence

.changeset/nice-weeks-enter.md

@@ -0,0 +1,5 @@
+---
+"unicode-segmenter": patch
+---
+
+Fix `Extend + Extended_Pictographic` cluster break

README.md

@@ -215,7 +215,7 @@ Since [Hermes doesn't support the `Intl.Segmenter` API](https://github.com/faceb
 
 | Name                         | Unicode® | ESM? |   Size    | Size (min) | Size (min+gzip) | Size (min+br) | Size (min+zstd) |
 |------------------------------|----------|------|----------:|-----------:|----------------:|--------------:|----------------:|
-| `unicode-segmenter/grapheme` |   17.0.0 |   ✔️ |    11,935 |      7,781 |           3,868 |         3,118 |           3,997 |
+| `unicode-segmenter/grapheme` |   17.0.0 |   ✔️ |    11,964 |      7,780 |           3,872 |         3,102 |           3,996 |
 | `graphemer`                  |   15.0.0 |   ✖️ ️|   410,435 |     95,104 |          15,752 |        10,660 |          15,911 |
 | `grapheme-splittetr`         |   10.0.0 |   ✖️ |   122,254 |     23,682 |           7,852 |         4,802 |           6,753 |
 | `@formatjs/intl-segmenter`*  |   17.0.0 |   ✖️ |   268,301 |    176,759 |          45,988 |        31,701 |          45,370 |
@@ -231,7 +231,7 @@ Since [Hermes doesn't support the `Intl.Segmenter` API](https://github.com/faceb
 
 | Name                         | Bytecode size | Bytecode size (gzip)* |
 |------------------------------|--------------:|----------------------:|
-| `unicode-segmenter/grapheme` |        21,427 |                12,048 |
+| `unicode-segmenter/grapheme` |        21,419 |                12,075 |
 | `graphemer`                  |       134,085 |                31,770 |
 | `grapheme-splitter`          |        63,942 |                19,165 |
 | `@formatjs/intl-segmenter`   |       329,547 |               136,751 |

src/grapheme.js

@@ -65,9 +65,12 @@ export function* graphemeSegments(input) {
   /** The number of RI codepoints preceding `cursor`. */
   let riCount = 0;
 
+  /** Tracks if Extended_Pictographic was seen in the current Extend* sequence for GB11 */
+  let extPic = catBefore === 4;
+
   /**
    * Emoji state for GB11: tracks if we've seen Extended_Pictographic followed by Extend* ZWJ
-   * Only relevant when catBefore === ZWJ && catAfter === Extended_Pictographic
+   * Only relevant when catBefore === ZWJ && extPic (catAfter === Extended_Pictographic)
    */
   let emoji = false;
 
@@ -112,7 +115,7 @@ export function* graphemeSegments(input) {
       boundary = false;
     }
     // GB11: ExtPic Extend* ZWJ × ExtPic
-    else if (catBefore === 14 && catAfter === 4) {
+    else if (catBefore === 14 && extPic) {
       boundary = !emoji;
     }
     // GB12, GB13: RI × RI (odd count means no break)
@@ -149,6 +152,7 @@ export function* graphemeSegments(input) {
       };
 
       // Reset segment state
+      extPic = catAfter === 4;
       emoji = false;
       consonant = false;
       riCount = 0;
@@ -158,8 +162,8 @@ export function* graphemeSegments(input) {
     } 
     // Update state for continuing segment
     else {
-      // emoji state for GB11
-      if (catAfter === 14 && (catBefore === 3 || catBefore === 4)) {
+      // emoji state for GB11: ExtPic Extend* ZWJ × ExtPic
+      if (catAfter === 14 && extPic) {
         emoji = true;
       }
       // InCB state for GB9c

test/grapheme.js

@@ -308,6 +308,7 @@ test('counterexamples', async t => {
     ' କା',
     ' ଶ୍ୟା',
     'ക് വ',
+    'à‍❤',
   ];
 
   for (let counter of counterExamples) {