v0.3.24

Linux-X64: pandoc-crossref v0.3.24 git commit d84f1ed (HEAD) built with Pandoc v3.9, pandoc-types v1.23.1.1 and GHC 9.8.4
Windows-X64: pandoc-crossref v0.3.24 git commit d84f1ed (HEAD) built with Pandoc v3.9, pandoc-types v1.23.1.1 and GHC 9.8.4
macOS-ARM64: pandoc-crossref v0.3.24 git commit d84f1ed (HEAD) built with Pandoc v3.9, pandoc-types v1.23.1.1 and GHC 9.8.4
macOS-X64: pandoc-crossref v0.3.24 git commit d84f1ed (HEAD) built with Pandoc v3.9, pandoc-types v1.23.1.1 and GHC 9.8.4

Changelog

• Fix for #506

  A prior fix/workaround for jgm/pandoc#9720 had
  unintentionally clobbered explicitly-specified alt attributes on images.

  Unfortunately, Pandoc does not expose alt attribute explicitly to filters,
  instead it's only seen as image caption. So a simple heuristic is used:
  if image caption does not match the figure caption, we assume it's an
  explicitly-specified alt text and preserve it. Otherwise, clobber it.

• preserveAltText option. As stated above, if alt is missing or matches the
  figure caption, it will be clobbered, and figure caption will be generated
  by Pandoc with aria-hidden="true". When preserveAltText is set to True,
  pandoc-crossref will preserve the alt text (be it explicit or implicit from
  caption) verbatim, forcing Pandoc to produce both alt text and figure
  caption in the output. The hope is this may prove useful in some workflows,
  but you definitely should just specify alt text on all your images if you
  care about screen readers at all.