Fix some conversion errors on non conform PDF

[pubpub-zz]

tracked in #925

Iss due to an error in the stream pointing onto \n before the object

fix: move to next non whitespace

[codecov]

Codecov Report

Merging #932 (3e60f4c) into main (f261bad) will increase coverage by 4.50%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #932      +/-   ##
==========================================
+ Coverage   78.32%   82.83%   +4.50%     
==========================================
  Files          16       17       +1     
  Lines        4342     4106     -236     
  Branches      821      882      +61     
==========================================
  Hits         3401     3401              
+ Misses        754      515     -239     
- Partials      187      190       +3     

Impacted Files	Coverage Δ
PyPDF2/_reader.py	81.71% <100.00%> (+4.55%)	⬆️
PyPDF2/_page.py	71.70% <0.00%> (-4.74%)	⬇️
PyPDF2/_adobe_glyphs.py	100.00% <0.00%> (ø)
PyPDF2/xmp.py	60.86% <0.00%> (+0.24%)	⬆️
PyPDF2/_utils.py	91.52% <0.00%> (+0.84%)	⬆️
PyPDF2/_merger.py	67.90% <0.00%> (+2.29%)	⬆️
PyPDF2/_security.py	97.40% <0.00%> (+2.73%)	⬆️
PyPDF2/_writer.py	88.00% <0.00%> (+8.98%)	⬆️
PyPDF2/generic.py	89.82% <0.00%> (+12.05%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ba6614d...3e60f4c. Read the comment docs.

[MartinThoma]

@pubpub-zz Just a short heads-up: The reason why I didn't merge so far is copyright concerns. I have contacted the email mentioned in the PDF and asked for permission to add it to PyPDF2. If that permission is not given, we cannot add it.

Without the permission, I see three ways to merge the PR:

1. Removing the PDF (and the test, sadly)
2. Removing the PDF, but adding the PDF download within the test (instead of relying on the PDF being in the repo)
3. Removing the PDF, commenting the Test out, pointing to the PDF

[MartinThoma]

I fear that we will not get an answer and thus we will not be able to add this PDF.

@pubpub-zz I've tred to craft a PDF with that specific problem, but I failed. This is what I had:

    pdf_data = (
        b"%%PDF-1.7\n"
        b"1 0 obj << /Count 1 /Kids [4 0 R] /Type /Pages >> endobj\n"
        b"2 0 obj << >> endobj\n"
        b"3 0 obj << >> endobj\n"
        b"4 0 obj << /Contents 3 0 R /CropBox [0.0 0.0 2550.0 3508.0]"
        b" /MediaBox [0.0 0.0 2550.0 3508.0] /Parent 1 0 R"
        b" /Resources << /Font << >> >>"
        b" /Rotate 0 /Type /Page >> endobj\n"
        b"5 0 obj << /Pages 1 0 R /Type /Catalog >> endobj\n"
        b"xref 1 5\n"
        b"%010d 00000 n\n"
        b"%010d 00000 n\n"
        b"%010d 00000 n\n"
        b"%010d 00000 n\n"
        b"%010d 00000 n\n"
        b"trailer << /Root 5 0 R /Size 6 >>\n"
        b"startxref\n%d\n"
        b"%%%%EOF"
    )
    startx_correction = -1
    pdf_data = pdf_data % (
        pdf_data.find(b"1 0 obj") + startx_correction,
        pdf_data.find(b"2 0 obj") + startx_correction,
        pdf_data.find(b"3 0 obj") + startx_correction,
        pdf_data.find(b"4 0 obj") + startx_correction,
        pdf_data.find(b"5 0 obj") + startx_correction,
        # startx_correction should be -1 due to double % at the beginning indiducing an error on startxref computation
        pdf_data.find(b"xref") + startx_correction,
    )
    pdf_stream = io.BytesIO(pdf_data)
    reader = PdfReader(pdf_stream)

[pubpub-zz]

This is more tricky to reproduce: the fix deals with Objects stream. we can produce like that. :(
I've started to try to use pdftk but I think it will be difficult too...
we should focus on option 2...

[pubpub-zz]

@MartinThoma, ready for merging 😉

[MartinThoma]

@pubpub-zz Could you please run git rm resources/iss_925.pdf? The file is still in the PR and I cannot remove it without creating a new PR.