doc.close() not closing the document

[JojoMueller]

Description of the bug

Hi,

doc.close() does not really close a pdf file. My python script does some checks using PyMuPDF, closes the file, compresses it by calling an external program and finally saves the original and copies the compressed file onto the original one. This yields an os error, indicating that the original file is in use by the current python process. Importing fitz_old as fitz instead of fitz works fine.

Thank you in advance for your effort fixing the problem and enhancing this excellent program.

How to reproduce the bug

• Open a PDF file using PyMuPDF: doc = fitz.open(pdf_path)
• Try to close the document: doc.close()
• Try to copy a file onto the one pointed to by pdf_path: os.rename(any_file, pdf_path)

PyMuPDF version

1.23.17

Operating system

Windows

Python version

3.10

[julian-smith-artifex-com]

With rebased, the file is kept open after .close() if a fitz.Page instance is still alive.

So for example:

document = fitz.open(path)
page = document[0]
document.close()
# We still have a file handle open for `path`.
page = None
# No file handle open for `path`.

Arguably this new behaviour is ok, but we're looking at whether we could replicate the behaviour of fitz_old where page would become some sort of zombie object after document.close().

[JojoMueller]

Thanks for your help. After setting page to None the file is effectively closed.

[julian-smith-artifex-com]

Fixed in 1.23.18. (There is now no need to set page to None.)