PDF chapter split, reorder and bookmark review / PDF 章节拆分重排核对

All input PDFs are existing self-authored public samples. On 2026-10-13, the current online tools processed them through actual browser UI and the files were saved using actual download buttons.
输入均为既有自编公开样例。2026-10-13 通过当前线上工具实际选择文件、处理并点击下载；输出文件原始字节未改写。

chapters-source.pdf: eight pages, two front-matter pages and three two-page chapters. Top-level bookmark targets are file pages 3,5,7; children target 4,6,8.
split-current.zip: native Download all result, three complete chapters. Each PDF has two pages and a two-level bookmark tree pointing to local pages 1,2.
first-chapter-current.pdf: native individual download; exactly equals part-001.pdf inside split-current.zip.
chapters-review-then-area.pdf: native sequential merge of part-003.pdf followed by part-001.pdf, Preserve bookmarks enabled. Output is four pages corresponding to original file pages 7,8,3,4. Four retained bookmark targets are 1,2,3,4 in the new document, in Chapter3/child/Chapter1/child tree order.

合并没有重写正文里的 FILE PAGE 标记、印刷页码或章节标题。新文件第1页仍含原文件第7页的标记和印刷第5页，不能因此把书签改回7。四条书签已在书签工具重新打开后逐条点击，实际页面指示器分别为1/4、2/4、3/4、4/4。
The merge does not rewrite printed page numbers, original FILE PAGE labels or chapter titles. New page1 still carries original file-page7 and printed-page5. Do not reset its bookmark to7. All four bookmarks were clicked after reopening the actual merged download, reaching 1/4 through4/4.

source-a.pdf / source-b.pdf: separate 3-page and2-page labelled sources used to recheck the old two-page-interleave limitation.
block-two-current.pdf: native current output using Interleaved merge, two pages per file, Preserve bookmarks enabled. Page order is A1,A2,B1,B2,A3. Bookmark-tree destinations for A1,A2,A3,B1,B2 are1,2,5,3,4; tree order follows source grouping and is not a sorted page-number list.

checks.json: page text, dimensions, rotation and outline checks, plus observed bookmark-click results.
This ZIP is assembled locally for comparison; its contained native outputs are unchanged. No OCR, translation or AI bookmark generation was used. These small text PDFs do not verify damaged/encrypted/signed files, all viewers or large-document limits.
本ZIP为本地整理核对包，包内原生输出不变。没有OCR、翻译或AI目录识别；小型文字样例不能证明损坏/加密/签名文件、所有阅读器或大型文档表现。
