Codecov Report

Base: 91.92% // Head: 91.92% // Decreases project coverage by -0.01% ⚠️

Coverage data is based on head (26bdba2) compared to base (4e276b2).
Patch coverage: 100.00% of modified lines in pull request are covered.

@@            Coverage Diff             @@
##             main    #1639      +/-   ##
==========================================
- Coverage   91.92%   91.92%   -0.01%     
==========================================
  Files          33       33              
  Lines        6375     6374       -1     
  Branches     1272     1272              
==========================================
- Hits         5860     5859       -1     
  Misses        327      327              
  Partials      188      188              

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

Nevermind, doesn't quite work yet, seemingly only renders the first page correctly in Adobe Acrobat.

I'll reopen this, because this does seem to fix /Root/AcroForm getting generated incorrectly and thus the proper setting of the /NeedAppearances flag. With this at least, it's possible to render the first page correctly in Adobe Reader (since Adobe Reader has a bug that only respects the flag for the first page of a document) and all pages in most other PDF readers.

@cryzed,
can you provide a pdf that requires this change: I do not mean the original code is correct but I would have had expected the /AcroFrom being already correcty genenated as the /Fields should be correct

Get_Started_With_Smallpdf.pdf

# Taken from https://pypdf.readthedocs.io/en/stable/user/forms.html
from pypdf import PdfReader, PdfWriter

reader = PdfReader("Get_Started_With_Smallpdf.pdf")
writer = PdfWriter()

page = reader.pages[0]
fields = reader.get_fields()

writer.add_page(page)

writer.update_page_form_field_values(
    writer.pages[0], {"test_field": "some filled in text"}
)

# write "output" to pypdf-output.pdf
with open("filled-out.pdf", "wb") as output_stream:
    writer.write(output_stream)

PdfWriter.set_need_appearances_writer() is called implicitly by PdfWriter.update_page_form_field_values(). Open filled-out.pdf in Adobe Reader and it won't render any appearances, even on the first page. Same thing happens in Okular.

add_page() is a quite low level function to add page. I recomment to use .append() which is easier to use and more powerfull.

import pyppdf

w = pypdf.PdfWriter()
w.append("filledGet_Started_With_Smallpdf.pdf", [0])
w.update_page_form_field_values(w.pages[0], {"test_field": "some filled in text"})
w.write("tt.pdf")

Although your change is a good idea 😊

This patch does work for me in Adobe Acrobat (2022.003.20322), PDF-XChange and Edge browser but does not work in Chrome or Vivaldi browser.

@pubpub-zz Do you think this PR improves pypdf? Or is there an issue we need to address before merging?

the new code is better than the old one. and may explain why Acrobat reader may have some difficulties to read some files. For me you can merge it.

@cryzed Thank you for your PR 🙏

I'm sorry it took so long until it was merged. I misunderstood a part of the conversation and thought there was an issue with the PR + I had some health issues (luckily all good again :-) )

Your fix will be part of the next release, likely on 12.05.2023 with pypdf>3.5.1

If you want, I can add you to https://pypdf.readthedocs.io/en/latest/meta/CONTRIBUTORS.html :-)

If you want, I can add you to https://pypdf.readthedocs.io/en/latest/meta/CONTRIBUTORS.html :-)

Sure, sounds good to me, if you think this was an adequate contribution!