jgm/pandoc: pandoc 2.13

Click to expand changelog - Support `yaml_metadata_block` extension for `commonmark`, `gfm` (#6537). This supported is a bit more limited than with pandoc's `markdown`. The YAML block must be the first thing in the input, and the leaf notes are parsed in isolation from the rest of the document. So, for example, you can't use reference links if the references are defined later in the document. - Fix fallback to default partials when custom templates are used. If the directory containing a template does not contain the partial, it should be sought in the default templates, but this was not working properly (#7164). - Handle `nocite` better with `--biblatex` and `--natbib` (#4585). Previously the nocite metadata field was ignored with these formats. Now it populates a `nocite-ids` template variable and causes a `\nocite` command to be issued. - Text.Pandoc.Citeproc: apply `fixLinks` correctly (#7130). This is code that incorporates a prefix like `https://doi.org/` into a following link when appropriate. - Text.Pandoc.Shared: - Remove `backslashEscapes`, `escapeStringUsing` \[API change\]. Replace these inefficient association list lookups with more efficient escaping functions in the writers that used them (for a 10-25% performance boost in org, haddock, rtf, texinfo writers). - Remove `ToString`, `ToText` typeclasses \[API change\]. These were needed for the transition from String to Text, but they are no longer used and may clash with other things. - Simplify `compactDL`. - Text.Pandoc.Parsing: - Change type of `readWithM` so that it is no longer polymorphic \[API change\]. The `ToText` class has been removed, and now that we've completed the transition to Text we no longer need this to operate on Strings. - Remove `F` type synonym \[API change\]. Muse and Org were defining their own `F` anyway. - Text.Pandoc.Readers.Metadata: - Export `yamlMetaBlock` \[API change\]. - Make `yamlBsToMeta`, `yamlBsToRefs` polymorphic on the parser state \[API change\]. - Markdown reader: Fix regression with `tex_math_backslash` (#7155). - MediaWiki reader: Allow block-level content in notes (ref) (#7145). - Jira reader (Albert Krewinkel): - Fixed parsing of autolinks (i.e., of bare URLs in the text). Previously an autolink would take up the rest of a line, as spaces were allowed characters in these items. - Emoji character sequences no longer cause parsing failures. This was due to missing backtracking when emoji parsing fails. - Mark divs created from panels with class "panel". - RST reader: fix logic for ending comments (#7134). Previously comments sometimes got extended too far. - DocBook writer: include Header attributes as XML attributes on section (Erik Rask). Attributes with key names that are not allowed as XML attributes are dropped, as are attributes with invalid values and `xml:id` (DocBook 5) and `id` (DocBook 4). - Docx writer: - Make `nsid` in `abstractNum` deterministic. Previously we assigned a random number, but we don't need random values, so now we just assign a value based on the list marker. - Use integral values for `w:tblW` (#7141). - Jira writer (Albert Krewinkel): - Block quotes are only rendered as `bq.` if they do not contain a linebreak. - Jira writer: improve div/panel handling. Include div attributes in panels, always render divs with class `panel` as panels, and avoid nesting of panels. - HTML writer: Add warnings on duplicate attribute values. This prevents emitting invalid HTML. Ultimately it would be good to prevent this in the types themselves, but this is better for now. - Org writer: Prevent unintended creation of ordered list items (#7132, Albert Krewinkel). Adjust line wrapping if default wrapping would cause a line to be read as an ordered list item. - JATS templates: support 'equal-contrib' attrib for authors (Albert Krewinkel). Authors who contributed equally to a paper may be marked with `equal-contrib`. - reveal.js template: replace JS comment with HTML (#7154, Florian Kohrt). - Text.Pandoc.Logging: Add `DuplicateAttribute` constructor to `LogMessage`. \[API change\] - Use `-j4` for linux release build. This speeds up the build dramatically on arm. - cabal.project: remove ghcoptions. Move flags to top level, so they can be set differently on the command line. - Require latest texmath, skylighting, citeproc, jira-wiki-markup. (The latest skylighting fixes a bad bug with Haskell syntax highlighting.) Narrow version bounds for texmath, skylighting, and citeproc, since the test output depend on them. - Use doclayout 0.3.0.2. This significantly reduces the time and memory needed to compile pandoc. - Use `foldl'` instead of `foldl` everywhere. - Update bounds for random (#7156, Alexey Kuleshevich). - Remove uses of some partial functions. - Don't bake in a larger stack size for the executable. - Test improvements: - Use `getExecutablePath` from base, avoiding the dependency on `executable-path`. - Factor out `setupEnvironment` in Helpers, to avoid code duplication. - Fix finding of data files by setting teh `pandoc_datadir` environment variable when we shell out to pandoc. This avoids the need to use `--data-dir` for the tests, which caused problems finding `pandoc.lua` when compiling without the `embed_data_files` flag (#7163). - Benchmark improvements: - Build `+RTS -A8m -RTS` into default ghc-options for benchmark. This is necessary to get accurate benchmark results; otherwise we are largely measuring garbage collecting, some not related to the current benchmark. - Allow specifying BASELINE file in 'make bench' for comparison (otherwise the latest benchmark is chosen by default). - Force `readFile` in benchmarks early (Bodigrim). - CONTRIBUTING: suggest using a `cabal.project.local` file (#7153, Albert Krewinkel). - Add ghcid-test to Makefile. This loads the test suite in ghcid