Possible regression in pdf cleaning during save.

[wz93672]

Description of the bug

Recently, in the documents to process, I received a document scanned with a Lexmark machine that become blank after saving with “clean” option set to True. This behavior starts with version 1.24.0, 1.23.26 works fine.

How to reproduce the bug

Here is a public document scanned with Lexmark, found in the internet:
https://www.feb.unesp.br/Home/Administracao110/DTAd/Compras/empenhos---19_06.pdf

with pymupdf.open('empenhos---19_06.pdf') as doc:
    doc.save('out.pdf', clean=True)

PyMuPDF version

1.24.13

Operating system

Windows

Python version

3.12

[JorjMcKie]

What are you trying to achieve with this parameter?
It is equivalent to executing page.clean_contents() for all its pages. For PDF as you used it for, it has no positive effect in any case. It never reduces files size if not used together with garbage collection and compression.
Of course what you experienced shouldn't happen either.

[JorjMcKie]

Here is the link to the corresponding MuPDF bug report: https://bugs.ghostscript.com/show_bug.cgi?id=708128

[wz93672]

It was only minimal example to reproduce the bug.
I’m using pymupdf for preprocessing mix of pdf documents form different sources, quiet a lot of files, including merging, splitting, adding blank pages for parity, to streamline printing. Some are generated in older software, or scanned on older machines with old firmware, and have problems. Among them there are pdfs with broken blank pages I wrote about a few months ago, or with broken coordinate system (unclosed transformations as I understand). Flawed files come and go, disappearing characters, nulls after eof. Typically works when opened directly but breaks on save, or with whole setup gives hard to adress problems. Cleaning helps and currently I have to use it, I wish cleaning would purge them all :). And I’m using garbage option too.

[julian-smith-artifex-com]

Fixed in PyMuPDF-1.25.1.