Signature coordinates error on different operating systems

[haumenphai]

I tested pyhanko on Window and Ubuntu,
I have a question why the coordinates of signature field are different.

for example with coordinates:

signature_coords = {'x1': 158.30325, 'y1': 591.17825, 'x2': 278.30325000000005, 'y2': 666.1782499999999}

on Ubuntu fill in the order x1,y1,x2,y2 and the result is as expected.

                pdf_signer = signers.PdfSigner(
                    signature_meta=signers.PdfSignatureMetadata(
                        field_name=data['signature_field_name']
                    ),
                    new_field_spec=SigFieldSpec(
                        sig_field_name=data['signature_field_name'],
                        on_page=data['on_page'] - 1,
                        box=(
                            int(signature_coords['x1']),
                            int(signature_coords['y1']),
                            int(signature_coords['x2']),
                            int(signature_coords['y2'])
                        )
                    ),
         
                    signer=window_signer
                )

But on Windows, I have to fill in differently to get the correct box like on Ubuntu.

box=(
    int(signature_coords['x1']),
    int(signature_coords['y2']),
    int(signature_coords['x2']),
    int(signature_coords['y1'])
)

I want my code to run well on different operating systems, this makes me wonder.

===
version: 0.25.1

[haumenphai]

Yes, I sloved that problem, but i'm not sure if it always happens with windows os.

[MatthiasValvekens]

@niranjanpathak-np1 While your attempt at a solution is appreciated, please don't just blindly post LLM-generated answers to questions. Especially when the LLM misses the point by a mile...

Anyway, this is not likely to be a pyHanko issue, but rather a "I have two PDF viewers that produce different results when fed the same file" issue. The difference in OSes is also not really relevant. PyHanko is completely platform-agnostic and performs the exact same PDF manipulations regardless of the platform on which it is executed.

Please include the names & versions of the viewers you're using, and also your input document, since this can only be pyHanko's fault if it somehow turned a valid PDF into an invalid one during the signing process.

I'll also be converting this into a discussion.