bug: incomplete redaction of vector graphics (line art) #134

mahlzahn commented 2026年03月12日 14:38:59 +01:00

Description

Although the "drawings" in the sample PDF attached, are not visible, they may in the worst case contain sensible information and should probably be removed, too, when applying redactions.

How to reproduce

With attached sample pdf file, do

from pymupdf import *
dx = Document("incomplete_redaction_sample.pdf")
dx[0].add_redact_annot(Rect(FZ_MIN_INF_RECT, FZ_MIN_INF_RECT, FZ_MAX_INF_RECT, FZ_MAX_INF_RECT), fill=False)
dx[0].apply_redactions(graphics=2)
dx[0].get_drawings()

This returns still 4 drawings ...

Originally, this was found on page 20 (i = 23) of https://github.com/py-pdf/sample-files/tree/main/009-pdflatex-geotopo.

Reported upstream issue

https://github.com/pymupdf/PyMuPDF/issues/4936
https://bugs.ghostscript.com/show_bug.cgi?id=709237

Impact

It seems to be a not so rare issue, it was found in 27 of approximately 1,000 PDF document samples, see #110.

## Description Although the "drawings" in the sample PDF attached, are not visible, they may in the worst case contain sensible information and should probably be removed, too, when applying redactions. ## How to reproduce With attached sample pdf file, do ```py from pymupdf import * dx = Document("incomplete_redaction_sample.pdf") dx[0].add_redact_annot(Rect(FZ_MIN_INF_RECT, FZ_MIN_INF_RECT, FZ_MAX_INF_RECT, FZ_MAX_INF_RECT), fill=False) dx[0].apply_redactions(graphics=2) dx[0].get_drawings() ``` This returns still 4 drawings ... Originally, this was found on page 20 ($i = 23$) of https://github.com/py-pdf/sample-files/tree/main/009-pdflatex-geotopo. ## Reported upstream issue https://github.com/pymupdf/PyMuPDF/issues/4936 https://bugs.ghostscript.com/show_bug.cgi?id=709237 ## Impact It seems to be a not so rare issue, it was found in 27 of approximately 1,000 PDF document samples, see #110.