v1.15.3

Bug fixes

• Preserve XML names and visible content, Markdown whitespace, delimiters, and image URLs, and source text across chunked input. Keep CSS minification stable across repeated passes, align sanitizer handling of nested options, and limit Markdown output growth for deeply nested inline markup. Raise MemoryError for allocation failures in DOM, CSS, XPath, and extraction APIs. (#1536) by @gaborbernat.
• Keep sanitizer string and node outputs aligned for select and nested options. Preserve nested Markdown emphasis, punctuation, and breaks across CommonMark reparses. Preserve CSS margin values when a shorthand falls between longhands, and keep ragged Markdown tables as HTML. Report allocation failures during text rendering, element text and textarea value assignment, microdata and feed extraction, and encoding detection. Preserve escaped commas and spacing before negative dimensions in CSS, and fold valid calc() expressions with comments in one pass. Bound parsing and sanitizing time on deeply nested lists, and avoid rescanning stray CSS preludes without a block. (#1537) by @gaborbernat.

🐍 PyPI release

📜 Full Changelog