v0.11.0

Added

• The content health page opens with an overview: pages on the site by kind,
  files not on it, notes and documents to fix, each linking to its section,
  and when the site last changed. "Not on the site" opens with a table of its
  groups — what, how many, where, what to do — before the lists. See
  ADR-028.
• Files the site does not publish are grouped by what to do with them: Word
  and text files (save as Google Docs), presentations (export as PDF),
  archives, images, diagrams, spreadsheets, video and audio, other files.
• Notes: pages that are on the site with something to know — an image or a
  link conversion left out, a link to a heading that now opens the top of a
  page, a merged cell split, a code block never closed, files in one folder
  with the same name, an order number used twice, an ignored folder that is
  not there.
• The job summary of a sync says what the run changed, page by page with
  links, and what stopped a run that failed, with what it means and what to
  do. The state of the site follows in the same grouped tables as the page.

Changed

• The sync report moves to schema version 3: both catalogs, the notes, pages
  by kind, and how the pages were made. A run counts pages it added, changed
  and removed by their output, and documents it exported separately, so a
  full sync over unchanged documents reports nothing changed.

Fixed

• The text of a PDF no longer runs words together where a slide or a table
  draws them as separate pieces, and a line in capitals is a paragraph of its
  own instead of running into the sentence before it. Every published PDF is
  read again once on the next sync.