fpdf2 2.8.6 on Python PyPI

[2.8.6] - 2026-02-18

support for FPDF.transform() context manager to apply geometric transformations (translation, rotation, scaling, skewing) to the current graphics state - thanks to @RaphaelFLT
support for SVG <linearGradient> and <radialGradient> elements - cf. issue #1580 - thanks to @Ani07-05
mypy and pyright checks in the CI pipeline to enforce strict typing
support WOFF and WOFF2 fonts - thanks to @BharathPESU
TTC/OTC face selection via collection_font_number in add_font()
CID-keyed CFF font embedding support
Microsoft Symbol font remapping for non-Unicode cmaps
support for EBDT/EBLC bitmap fonts

the A5 value that could be specified as page format to the FPDF constructor was slightly incorrect, and the corresponding page dimensions have been fixed. This could lead to a minor change in your documents dimensions if you used this A5 page format. - cf. issue #1699
a bug when rendering empty tables with INTERNAL layout, that caused an extra border to be rendered due to an erroneous use of list.index() - cf. issue #1669
edge cases where tiny floating-point drift could trigger incorrect behaviour like issue #1393
parentheses are now escaped in more places. This fixes embedding a file with a single parenthesis in the filename, but might also have impacts on e.g. section names.
parsing markdown sequences with escape character - cf. issue #1236 - thanks @amidou-naba

improved performance when rendering paths, SVGs, and opaque raster images with an alpha channel - cf. PR #1675
typing annotations added across the codebase as part of the strict typing rollout
graphics state snapshots now use a GraphicsState dataclass dataclass for clearer usage and stronger typing
rotation(), skew(), and mirror() now delegate to transform() for standardized transform application
ImageInfo now uses typed dictionaries for clearer usage and stronger typing