PyMuPDF-2.0 released

[julian-smith-artifex-com]

Wheels for Windows, Linux and MacOS, and the sdist, are available on
https://pypi.org and can be installed in the usual way, for example:

python -m pip install --upgrade pymupdf

Changes in version 2.0

Fixed issues:

• Fixed Dehyphenate does no longer work since 1.27.1 #4994
• Fixed clean_contents(sanitize=True) drops glyph positioning, duplicating an overprinted leading glyph (regression in 1.28.0 from 1.27.2.3) #5054
• Fixed insert_file(...) breaks radio button groups #5055
• Fixed Annotations disappear or move when using Page.remove_rotation() for 90 or 270 rotated pages #5059
• Fixed Memory leak in extractBLOCKS (_as_blocks): text object reference never released #5067
• Fixed Page.remove_rotation() takes minutes on pages with thousands of links #5077
• Fixed Regression in 1.28.x: get_pixmap() and get_text() hang at 100% CPU on a specific PDF; works in 1.27.2.3 #5082
• Fixed Import of fitz prints to console instead of generating a DeprecationWarning #5100
• Fixed Cyclic Widget /Parent makes page operations fail with FzErrorFormat: code=7: cycle in resources #5101
• Fixed Page.clip_to_rect() drops 'sh' (shading) operators #5108
• Fixed Stack-based buffer overflow in JM_set_annot_callout_line #5112
• Fixed get_drawings() reports a wrong start point for a subpath following a re with negative width/height #5122
• Fixed get_pixmap() never returns on a tagged PDF page with many MCIDs — pdf_lookup_mcid_in_mcids is a linear scan, making rendering O(n²) per page #5125
• Fixed Page.get_label() raises IndexError when the page precedes the first label definition #5140
• Fixed apply_redactions() does not remove text shown with the ' operator after TL + Td (regression in 1.28) #5152
• Fixed insert_pdf raises ValueError: <xref> is not in list when a widget's /Parent does not list it in /Kids #5154
• Fixed Page.add_widget fails for radio buttons #5165

Other:

• Added support for form field hierarchies.
• Added support for radio button groups.
• Added support for widget aliases.
• Fixed builds with swig-4.5.
• Fixed https://bugs.ghostscript.com/show_bug.cgi?id=709683_ Path Traversal and Arbitrary File Write in PyMuPDF CLI Font Extraction (CVE-2026-82035)
• Fixed ignoring aspect ratio for square images.
• Fixed potential SEGV in pymupdf.Page.get_cdrawings() if callback raises exception.
• Fixed use with python-3.15.
• Remove the debug timing loop from Annot.update().
• Retrospectively marked Page.delete_widget() doesn't fully remove the widget, other programs still detect the widgets #3478 fixed in 1.25.3.
• Retrospectively marked PyMuPDF Pro 1.25.0: Some graphical elements of a .docx file not visible with get_pixmap() #4158 fixed in 1.27.2.2
• Retrospectively marked write_images() and saving with garbage>1 spoils fonts (xrefs) #4896 fixed in 1.28.0.
• Retrospectively marked pdf_rewrite_images() segfaults with shared image xrefs across many pages (buffer overflow) #4918 fixed in 1.28.0.
• Supported Python versions are now 3.11-3.15.
• Use MuPDF-1.28.5.
• Warning for use of deprecated fitz API now uses warnings.warn() with FutureWarning.

---

This discussion was created from the release PyMuPDF-2.0 released.