[Bug]: Conversion to PDF/A strips Acrobat-produced CJK text layer

[nisbet-hubbard]

Describe the bug

We rely on adobe acrobat for CJK texts because of tesseract’s spacing issues (tesseract-ocr/tesseract#2702).

Currently, however, the pdf/a mode of ocrmypdf strips such an acrobat-produced pdf of its OCR layer.

Steps to reproduce

1. Run `ocrmypdf --skip-text input.pdf output.pdf`
2. `pdffonts input.pdf`
3. `pdffonts output.pdf`

The results of 2 and 3 are the same if `--output-type pdf` is used.

Files

No response

How did you download and install the software?

PyPI (pip, poetry, pipx, etc.)

OCRmyPDF version

16.10.1

Relevant log output