Debian Patches

Status for pikepdf/10.18.0-1

Patch Description Author Forwarded Bugs Origin Last update
drop-save-pike.patch Drop non-DFSG-compatible image from docs Sean Whitton <spwhitton@spwhitton.name> not-needed 2025-08-13
Optional-patch-remove-sphinx-design.patch Optional patch: remove sphinx-design
For Debian's pikepdf docs build.

(cherry picked from commit bfe76fdeb2d1a94b121493caacc7131a9232cb4f)
"James R. Barlow" <james@purplerock.ca> not-needed 2022-09-26
fix-test-only-pikepdf-pages.patch Fix "only pikepdf pages" __setitem__ test case
I observed the following test failure while updating Debian's pikepdf
package:

```
=================================== FAILURES ===================================
_______________ test_stress_extend_from_iterables_spanning_pdfs ________________
[gw2] linux -- Python 3.14.10 /usr/bin/python3

def test_stress_extend_from_iterables_spanning_pdfs():
"""extend() and slice assignment from iterables that run code as they go.

The iterable yields pages of several Pdfs and uses the destination while
it is being unpacked. An iterable with an item that is not a page raises
before any of its pages is inserted.
"""
sources = [pdf_with_pages(2, width=200 + 10 * k) for k in range(3)]
dst = pdf_with_pages(1, width=100)
rounds = 5

def pages_of_sources(i):
for source in sources:
assert len(dst.pages) >= 1 # Runs while the iterable is unpacked
yield source.pages[i % 2]

def extend(i):
for _ in range(rounds):
dst.pages.extend(pages_of_sources(i))
dst.pages[0:0] = pages_of_sources(i)
with pytest.raises(TypeError, match='only pikepdf pages'):
dst.pages.extend(itertools.chain(pages_of_sources(i), [42]))
with pytest.raises(TypeError, match='only pikepdf pages'):
dst.pages[0:0] = itertools.chain(pages_of_sources(i), [None])

def other_direction(k):
# Each source also takes pages from the destination, so that the
# locks of the same Pdfs are wanted in the opposite order.
for _ in range(rounds):
sources[k].pages.extend(page for page in dst.pages[0:1])
del sources[k].pages[-1]

> run_threads(
*[lambda i=i: extend(i) for i in range(4)],
*[lambda k=k: other_direction(k) for k in range(3)],
lambda: churn_pages(sources[0], 100),
)

tests/test_threading_pdf.py:216:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests/test_threading.py:351: in run_threads
raise errors[0]
tests/test_threading.py:329: in runner
target()
tests/test_threading_pdf.py:217: in <lambda>
*[lambda i=i: extend(i) for i in range(4)],
^^^^^^^^^
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

i = 0

def extend(i):
for _ in range(rounds):
dst.pages.extend(pages_of_sources(i))
dst.pages[0:0] = pages_of_sources(i)
with pytest.raises(TypeError, match='only pikepdf pages'):
dst.pages.extend(itertools.chain(pages_of_sources(i), [42]))
> with pytest.raises(TypeError, match='only pikepdf pages'):
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
E Failed: DID NOT RAISE TypeError

tests/test_threading_pdf.py:206: Failed
```

I can't reproduce this with `uv run -p 3.14 pytest
tests/test_threading_pdf.py`, so I'm not completely sure I've understood
the problem correctly. However, this test seems to be exercising both
`extend` and `__setitem__`, and appears to be trying to do so in the
same way earlier, so it seems that the arguments in both cases ought to
match.
Colin Watson <cjwatson@debian.org> yes 2026-11-10

All known versions for source package 'pikepdf'

Links