Entity handling fixes:

- Text.Pandoc.XML.fromEntities:  handle entities without a
  semicolon. Always lookup character references with the
  trailing ';', even if it wasn't present.  And never add
  it when looking up numerical entities.  (This is what
  tagsoup seems to require.)
- Text.Pandoc.Parsing.characterReference:  Always lookup
  character references with the trailing ';', and leave off
  the ';' when looking up numerical entities.

This fixes a regression for e.g. `⟨`.

0 files changed, 0 insertions, 0 deletions