#1140397 diffoscope: Diffoscope fails to compare PDF files - dumppdf crash

#1140397#5
Date:
2026-06-19 14:18:09 UTC
From:
To:
Dear Maintainer,

See this salsa build job comparing the "allabout.pdf" file:
https://nikolas.pages.debian.net/-/gnubg/-/jobs/9808951/artifacts/debian/output/diffoscope.html

diffoscope emits this message:

./usr/share/doc/gnubg/allabout.pdf
Command `'dumppdf -at {}'` failed with exit code 1. Standard output:
    Traceback (most recent call last):
      File "/usr/bin/dumppdf", line 468, in <module>
        main()
        ~~~~^^
      File "/usr/bin/dumppdf", line 454, in main
        dumppdf(
        ~~~~~~~^
            outfp,
          [...]

after installing the following packages in the debrebuild job on salsa:
  poppler-utils python3-pdfminer xxd

#1140397#10
Date:
2026-06-20 03:10:39 UTC
From:
To:
This appears to be a pdfminer bug. The fix is here:
https://github.com/pdfminer/pdfminer.six/pull/1270

#1140397#15
Date:
2026-06-22 19:49:23 UTC
From:
To:
Hey Nikolas,

Thanks for tracking this down and submitting a fix. :)  I'll let the
pdfminer folks fix this in "their" package rather than applying any
kind of workaround to diffoscope. 👍


Best wishes,

#1140397#30
Date:
2026-07-13 15:30:08 UTC
From:
To:
We can consider making a debian patch for pdfminer for now - upstream
has not moved yet, but the fix is quite straightforward:

https://github.com/pdfminer/pdfminer.six/pull/1271/changes

I will maybe come up with the merge request in salsa with these changes
if that's helpful.

#1140397#35
Date:
2026-07-15 17:15:30 UTC
From:
To:
Hello Nikolas,
Thanks for your report and sorry for my late reply. I hope to be able to
work on this on Sunday.
No need to come up with a merge request if you don't have time right
now. I hope to upload it on Sunday.

Cheers,
Daniele

#1140397#40
Date:
2026-07-21 00:37:07 UTC
From:
To:
Hello,
Here is a merge request with a patch for this:

https://salsa.debian.org/python-team/packages/pdfminer/-/merge_requests/3