Method getWordBoundaries

src/utils/Cursor.ts:1173–1189

* Get word boundaries using Intl.Segmenter for proper Unicode word segmentation. * This correctly handles CJK (Chinese, Japanese, Korean) text where each character * is typically its own word, as well as scripts that use spaces between words.

()

   * is typically its own word, as well as scripts that use spaces between words.
   */
  public getWordBoundaries(): Array<{
    start: number
    end: number
    isWordLike: boolean
  }> {
    if (!this.wordBoundariesCache) {
      this.wordBoundariesCache = []
      for (const segment of getWordSegmenter().segment(this.text)) {
        this.wordBoundariesCache.push({
          start: segment.index,
          end: segment.index + segment.segment.length,
          isWordLike: segment.isWordLike ?? false,
        })
      }
    }
    return this.wordBoundariesCache
  }

  /**
   * Binary search for boundaries.