fpdf2 Files

[2.8.8] - 2026-08-09

[!WARNING] Upgrading to fpdf2 2.8.8 is strongly recommended for applications that render user-provided images, SVGs, or HTML content.

Added

• Punjabi (pa) tutorial translation - thanks to @Pawansingh3889
• FPDF.svg_limits and SVGLimits to configure SVG complexity limits while rendering SVG images
• resource_access_policy and Security considerations documentation
• FPDF.optional_content() context manager to mark content as visible on screen only or in print only, using PDF Optional Content Groups - cf. issue [#441], based on a recipe by @digidigital - thanks to @eugen-goebel
• basic support for SVG <symbol> elements in the SVG parser - thanks to @Theo1335
• basic support for SVG <switch> elements in the SVG parser - cf. issue [#537] - thanks to @dannymaaz
• support for keeping aspect ratio for images in templates - cf. issue [#1118] - thanks to @prateek-dagar

Fixed

• custom height passed to Paragraph.ln() in a text region is now applied to the line it terminates instead of the first line of the following paragraph - cf. issue [#1786] - thanks to @Sanjays2402
• the optional numpy import in image_parsing.py no longer crashes on CPUs unsupported by numpy's manylinux wheel baseline; RuntimeError is now treated the same as ImportError, so numpy degrades to unavailable instead of taking down import fpdf - cf. issue [#1908] - thanks to @stumpylog
• font state (family, style, size, current font, and the page-level "font is set" flag) no longer leaks back onto the FPDF instance after a text_columns() / text_region() context exits, so a subsequent pdf.cell() / pdf.write() renders at the caller's font instead of the last paragraph's - cf. issue [#1804] - thanks to @Pawansingh3889
• text rendering when the first text on a page starts with a fallback glyph - cf. issue [#1772]
• preserve boundary-neutral formatting during bidirectional text preprocessing - cf. issue [#1779]
• transform application on user space gradients - cf. issue [#1784]
• dependency extras for camelot-py and endesive on pyproject.toml - cf. issue [#1792]
• preserve link annotations during dry-run of FPDF.multi_cell - cf. issue [#1807] - thanks to @CoLa5
• preserve two consecutive markdown links (without space inbetween) - cf. issue [#1814] - thanks to @CoLa5
• support markdown style around markdown links - cf. issue [#1826] - thanks to @CoLa5
• Reset gstate for ToC-rendering - cf. issue [#1837] - thanks to @CoLa5
• preserve markdown format in FPDF.multi_cell in dry-run - cf. issue [#1840] - thanks to @CoLa5
• fix page order after dry-run of FPDF.multi_cell in ToC - cf. issue [#1836] - thanks to @CoLa5
• rendering SVG arcs with very small sweeps that previously rounded to zero - cf. issue [#1831]
• spurious "Not enough horizontal space to render a single character" error when text without break opportunities is split into many small fragments, e.g. by a fallback font alternating with the main font - cf. issue [#1250] - thanks to @uttam12331
• number of surviving escape characters - cf. issue [#1215] - thanks to @amidou-naba
• leading spaces on new lines inside <pre> and <pre><code> blocks are no longer dropped - cf. issue [#1063] - thanks to @eugen-goebel
• FPDF.set_font() can restore current_font when the selected font state diverged - cf. PR [#1872] - thanks to @gaoflow
• embed CID-keyed CFF fonts as raw CFF programs so browser PDF viewers render them correctly - cf. issue [#1874]
• fixed broken links on documentation not directly leading to the API reference - cf. issue [#1876] - thanks to @iamfazakb
• reject SVG <use> cycles and excessive nested expansion to prevent resource exhaustion in FPDF.image()
• count SVG <switch> elements in SVG complexity limits
• declare the default base state and display order for Optional Content Groups so PDF viewers can list layers correctly - cf. issue [#1895]

Changed

• skip byte-for-byte compressed data comparison when zlib-ng is detected, regardless of OS