fix: avoid quadratic repeated punctuation merges

Author: chenglou

src/analysis.ts

@@ -352,12 +352,14 @@ function splitTrailingForwardStickyCluster(text: string): { head: string, tail:
   }
 }
 
-function isRepeatedSingleCharRun(segment: string, ch: string): boolean {
-  if (segment.length === 0) return false
-  for (const part of segment) {
-    if (part !== ch) return false
-  }
-  return true
+function getRepeatableSingleCharRunChar(
+  text: string,
+  isWordLike: boolean,
+  kind: SegmentBreakKind,
+): string | null {
+  return kind === 'text' && !isWordLike && text.length === 1 && text !== '-' && text !== '—'
+    ? text
+    : null
 }
 
 function endsWithArabicNoSpacePunctuation(segment: string): boolean {
@@ -883,10 +885,14 @@ function buildMergedSegmentation(
   const mergedWordLike: boolean[] = []
   const mergedKinds: SegmentBreakKind[] = []
   const mergedStarts: number[] = []
+  // Track repeatable single-char punctuation runs structurally so identical
+  // merges stay O(1) instead of re-scanning the accumulated segment each time.
+  const mergedSingleCharRunChars: (string | null)[] = []
 
   for (const s of wordSegmenter.segment(normalized)) {
     for (const piece of splitSegmentByBreakKind(s.segment, s.isWordLike ?? false, s.index, whiteSpaceProfile)) {
       const isText = piece.kind === 'text'
+      const repeatableSingleCharRunChar = getRepeatableSingleCharRunChar(piece.text, piece.isWordLike, piece.kind)
 
       // First-pass keeps: no-space script-specific joins and punctuation glue
       // that depend on the immediately preceding text run.
@@ -901,6 +907,7 @@ function buildMergedSegmentation(
       ) {
         mergedTexts[mergedLen - 1] += piece.text
         mergedWordLike[mergedLen - 1] = mergedWordLike[mergedLen - 1]! || piece.isWordLike
+        mergedSingleCharRunChars[mergedLen - 1] = null
       } else if (
         isText &&
         mergedLen > 0 &&
@@ -910,6 +917,7 @@ function buildMergedSegmentation(
       ) {
         mergedTexts[mergedLen - 1] += piece.text
         mergedWordLike[mergedLen - 1] = mergedWordLike[mergedLen - 1]! || piece.isWordLike
+        mergedSingleCharRunChars[mergedLen - 1] = null
       } else if (
         isText &&
         mergedLen > 0 &&
@@ -918,6 +926,7 @@ function buildMergedSegmentation(
       ) {
         mergedTexts[mergedLen - 1] += piece.text
         mergedWordLike[mergedLen - 1] = mergedWordLike[mergedLen - 1]! || piece.isWordLike
+        mergedSingleCharRunChars[mergedLen - 1] = null
       } else if (
         isText &&
         mergedLen > 0 &&
@@ -928,15 +937,12 @@ function buildMergedSegmentation(
       ) {
         mergedTexts[mergedLen - 1] += piece.text
         mergedWordLike[mergedLen - 1] = true
+        mergedSingleCharRunChars[mergedLen - 1] = null
       } else if (
-        isText &&
-        !piece.isWordLike &&
+        repeatableSingleCharRunChar !== null &&
         mergedLen > 0 &&
         mergedKinds[mergedLen - 1] === 'text' &&
-        piece.text.length === 1 &&
-        piece.text !== '-' &&
-        piece.text !== '—' &&
-        isRepeatedSingleCharRun(mergedTexts[mergedLen - 1]!, piece.text)
+        mergedSingleCharRunChars[mergedLen - 1] === repeatableSingleCharRunChar
       ) {
         mergedTexts[mergedLen - 1] += piece.text
       } else if (
@@ -950,11 +956,13 @@ function buildMergedSegmentation(
         )
       ) {
         mergedTexts[mergedLen - 1] += piece.text
+        mergedSingleCharRunChars[mergedLen - 1] = null
       } else {
         mergedTexts[mergedLen] = piece.text
         mergedWordLike[mergedLen] = piece.isWordLike
         mergedKinds[mergedLen] = piece.kind
         mergedStarts[mergedLen] = piece.start
+        mergedSingleCharRunChars[mergedLen] = repeatableSingleCharRunChar
         mergedLen++
       }
     }

src/layout.test.ts

@@ -530,6 +530,12 @@ describe('prepare invariants', () => {
     expect(prepared.segments).toEqual(['===', ' ', 'heading', ' ', '==='])
   })
 
+  test('keeps long repeated punctuation runs coalesced', () => {
+    const text = '('.repeat(256)
+    const prepared = prepareWithSegments(text, FONT)
+    expect(prepared.segments).toEqual([text])
+  })
+
   test('applies CJK and Hangul punctuation attachment rules', () => {
     expect(prepareWithSegments('中文，测试。', FONT).segments).toEqual(['中', '文，', '测', '试。'])
     expect(prepareWithSegments('테스트입니다.', FONT).segments.at(-1)).toBe('다.')