` element or simply unwraps the Div if the wrapping element wouldn’t have any attributes. The new behavior gives better semantic results in most cases, as `
` format. The template loads highlight.js CSS, JS, and RevealHighlight plugin.\r\n\r\n- Markdown writer:\r\n\r\n - Ensure that `\\` line breaks are used for commonmark.\r\n - Use setext for headers containing line breaks for commonmark (#11341).\r\n - Allow display math to start/end with space (#11384). This reverts to earlier \\< 3.7 behavior.\r\n - Properly handle tables with footers (#11416).\r\n\r\n- JATS writer:\r\n\r\n - Fix XML output for nested figures (#11362, Albert Krewinkel). Subfigures are now wrapped inside a `` element. Furthermore, figure content that isn’t allowed as children of `` elements, such as raw text, gets wrapped in `` elements to ensure schema-conforming XML.\r\n\r\n- AsciiDoc writer:\r\n\r\n - Use doubled delims in more contexts (#11362). Also escape the `#` character.\r\n - Use a span with role for SmallCaps (#11374).\r\n - Export spaces inside delimited constructs like emph.\r\n\r\n- Docx writer:\r\n\r\n - Skip directory entries when building media overrides (#11379, You Jiangbin). Pandoc’s docx writer was previously adding an `` for `/word/media/` in `[Content_Types].xml` when the reference doc contains media, which violates OPC rules and causes Word to report corruption.\r\n - Refactor the monolithic `writeDocx` into a number of smaller functions (Claude Opus 4.5).\r\n - Replace generic XML traversal with direct path navigation (Claude Opus 4.5). Instead of using Data.Generics `everywhere` to traverse the entire XML tree when setting language attributes, navigate directly to the known path `w:docDefaults/w:rPr/w:lang`. This is more efficient and removes the dependency on `Data.Generics`.\r\n\r\n- EPUB writer:\r\n\r\n - Don’t use footnote backlinks for EPUBv3. Here we use aside elements, which are popups, and the backlinks are not needed; in some readers they cause a redundant number to appear, since the reader adds a note number.\r\n\r\n- MediaWiki writer:\r\n\r\n - Use Doc Text instead of Text for document construction (with Claude Open 4.5). This refactors the writer to use Text.DocLayout combinators (vcat, hcat, literal, blankline, cr, chomp) for building output, following the pattern used by other text format writers (RST, Markdown, Man). This enables better control over line spacing and paragraph separation.\r\n - Improve blank space around div elements (#11417). This is merely cosmetic.\r\n\r\n- PPTX writer:\r\n\r\n - Support notes field in metadata for title slide (#5844, Chris Callison-Burch). This adds support for a `notes` field in the YAML metadata block that will be used as speaker notes for the title slide in PowerPoint output. Previously, there was no way to add speaker notes to the title slide since it is generated from metadata rather than from content blocks.\r\n\r\n- LaTeX writer:\r\n\r\n - Add PDF standard support via DocumentMetadata (#11407, Gordon Woodhull with Claude Opus 4.5). The `pdfstandard` variable can be used to specify PDF standards (PDF/A, PDF/X, PDF/UA) in LaTeX output. This uses LaTeX’s `\\DocumentMetadata` command, which requires LuaLaTeX. PDF version requirements are automatically inferred, but can be explicitly overridden. Automatic tagging is added for standards that require it.\r\n\r\n- Typst template:\r\n\r\n - Fix keywords usage. (#11317, har7an).\r\n - Disable hyphenation for title, subtitle (#11375).\r\n - Improve accessibility of definition lists (#11436). Before this change, our show rule produced an error when typst was run with `--pdf-standard=ua1`.\r\n\r\n- HTML5 template:\r\n\r\n - Conditionally include lang attribute, instead of providing it with an empty value.\r\n\r\n- JATS template:\r\n\r\n - Fix author prefix placeholder (#11381, Christophe Dervieux).\r\n\r\n- Text.Pandoc.Citeproc:\r\n\r\n - Fix biblatex parsing of `@commentary` entries (#11322).\r\n - Fix typo affecting `jurisdiction` biblatex type (#11321).\r\n - Avoid adding an extra space at the beginning of a Cite.\r\n\r\n- Text.Pandoc.Shared:\r\n\r\n - Export `hasLineBreaks` \\[API change\\]. This was formerly defined in the DocBook writer but more generally useful.\r\n\r\n- Text.Pandoc.PDF:\r\n\r\n - Add `SOURCE_DATE_EPOCH` to verbose environment variable info.\r\n\r\n- Text.Pandoc.Class:\r\n\r\n - Factor out `openURL` into Text.Pandoc.Class.IO.HTTP (unexported module).\r\n\r\n- Lua subsystem (Albert Krewinkel):\r\n\r\n - Mark readers and writers with their types (#11367). The `pandoc.readers` and `pandoc.writers` maps now have string values instead of boolean values. The string signals the type of the reader/writer, `\"text\"` for TextReader/TextWriter and `\"bytestring\"` for ByteStringReader/ByteStringWriter.\r\n - Support equallity checks of LogMessage objects.\r\n - Add function `pandoc.with_state` (#10859). The function allows to run a callback with a modified pandoc state. This provides the ability to temporarily modify the resource path, the user data directory, and the HTTP request headers.\r\n - Let `pandoc.with_state` error on unknown options (#11376).\r\n - Add function `pandoc.utils.documentation` (#10999). This is now used to generate much of the Lua API documentation.\r\n\r\n- Text.Pandoc.App:\r\n\r\n - Remove redundant check for asciidoc in UnknownReader.\r\n\r\n- Text.Pandoc.Logging:\r\n\r\n - Add `pretty` field to ToJSON instance for LogMessage. This just reproduces the output of `logMessage`, for convenience for those who are using the JSON output outside of Haskell.\r\n\r\n- Text.Pandoc.Error:\r\n\r\n - Change PandocHttpError constructor to take a Text instead of an HttpException as the second argument (see #10980) \\[API change\\] Motivation: exposing HttpException in the public API makes it difficult to provide a version of pandoc that can be compiled to wasm, which currently can’t handle the network libraries.\r\n - Define `displayException` for PandocError. This is a behavior change, not an API change, since there was already a definition that defaulted to using Show. The change here is that we use `renderError` for a more human-readable version.\r\n\r\n- Drop support for compilation with GHC versions \\< 9.6.\r\n\r\n- Use released citeproc 0.13, djot 0.1.3, skylighting-format-blaze-html 0.1.2 (see #11423), texmath 0.13.1, asciidoc 0.1.0.1, typst-0.9, hslua 2.5.\r\n\r\n- Add `-Wno-deriving-typeble` to cabal ghc-options for ghc \\>= 9.12.\r\n\r\n- Add `http` cabal flag (#10980). This allows pandoc to be compiled without support for making HTTP requests, which is useful when WASM is the target.\r\n\r\n- pandoc-lua-engine has a new build flag `repl`, allowing support for the Lua repl to be disabled for the wasm build.\r\n\r\n- We now sign Windows artifacts with a code signing certificate provided by SignPath.\r\n\r\n- Mention Excel in cabal description.\r\n\r\n- Makefile: remove some obsolete targets.\r\n\r\n- MANUAL.txt:\r\n\r\n - Fix typo about `--chunk-template` (#11358, Albert Lei).\r\n - Fix link for bbcode_steam (#11389).\r\n - Small rewrite of syntax-highlighting info.\r\n - Fix defaults.yaml example for `wrap`.\r\n\r\n- Fix a few mistakes in the contributing docs. (#11318, har7an).\r\n\r\n- Fix a couple small errors in `doc/lua-filters.md`, `doc/custom-writers.md`, and `doc/custom-readers.md` (#11408, #11388).\r\n\r\n