Handling of incomplete EOF markers

[stefan6419846]

I just stumbled upon a PDF sent by e-mail where the final OF of the EOF marker has been missing. Apparently, pdf.js and Evince are able to display it correctly, while pypdf fails.

Environment

Which environment were you using when you encountered the problem?

$ python -m platform
Linux-5.14.21-150400.24.100-default-x86_64-with-glibc2.31

$ python -c "import pypdf;print(pypdf._debug_versions)"
pypdf==4.3.1, crypt_provider=('pycryptodome', '3.18.0'), PIL=10.0.0

Note: The version is the latest main code.

Code + PDF

This is a minimal, complete example that shows the issue:

from pypdf import PdfReader

reader = PdfReader('file.pdf')

For the file, just take any valid PDF file and remove the final OF characters of %%EOF.

Traceback

This is the complete traceback I see:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/stefan/temp/pdf/pypdf/_reader.py", line 142, in __init__
    self.read(stream)
  File "/home/stefan/temp/pdf/pypdf/_reader.py", line 566, in read
    self._find_eof_marker(stream)
  File "/home/stefan/temp/pdf/pypdf/_reader.py", line 658, in _find_eof_marker
    line = read_previous_line(stream)
  File "/home/stefan/temp/pdf/pypdf/_utils.py", line 279, in read_previous_line
    raise PdfStreamError(STREAM_TRUNCATED_PREMATURELY)
pypdf.errors.PdfStreamError: Stream has ended unexpectedly

[pubpub-zz]

there is one extra point behind that I've detected : If the %%EOF is truncated/missing, pypdf searches further in the file. If the file is linearized or has been incrementally built, a full "section" may be ignored without warning.