v0.8.0

• Now insists on Pandoc version 2.7.
• All files are now read and written in "UTF-8" encoding.
• Pushes now result in creation of a Docker image in Docker Hub, so there's no need to build a local Docker image.
• ./build-docker now just pulls the latest Docker image from Docker Hub.

v0.7.2

Another fix to Pandoc version check logic in build.

v0.7.0

Fixed pandoc version checking. Patch courtesy of @dnstanciu

Version 0.7.0

Courtesy of @szaffarano, there's now a ./build-docker script that will install and run the entire toolchain in an isolated Docker image. See this section in the README for details.

Version 0.6.1

Fixed Issue 1: The build wasn't ordering the chapters properly, because of a missing sort. Thanks to @szaffarano for the issue and the patch.

Version 0.6.0

• Added support for embedding PlantUML diagrams, provided the pandoc-plantuml-filter is installed.
• Added ability to create PDF using Weasy Print, rather than LaTeX, which can resolve font problems with some printers. Weasy Print must already be installed in your Python 3 environment, and it's not as simple as a pip install. See http://weasyprint.readthedocs.io/en/latest/install.html
• Added support for the following Pandoc extensions:
  • fenced_code_blocks
  • fenced_code_attributes
  • backtick_code_blocks
• Now ensures that generated HTML (including HTML used by Weasy Print) has the book language.
• Requires Pandoc 2.0.4 or better (and aborts if this requirement is not met).

Version 0.5.0

• Added ability to provide an author page (author.md).
• Added build code to find local image references in the book's Markdown sources, treating them as dependencies for the build.
• Cleaned up logic in scripts/pandoc-filter.py.
• Added some internal code documentation.
• Changed HTML styling to use traditional book-style paragraph indentation.
• upgrade.py now works even if the target directory is empty.

Version 0.4.0

• Added license terms (GPL v3).
• Added support for a (JavaScript-generated) table of contents in the HTML version of the book.
• Fixed table of contents generation with ePub. This task included (a) removing behavior in the Pandoc filter that short-circuited Pandoc's table of contents logic, and (b) adding some build code to rewrite the table of contents files to remove empty entries and entries that just pointed to title pages.
• Fixed center-, left- and right-justification logic in the filter to work if the paragraph is preceded by forced line breaks.
• Removed support for %newpage%. Just use an empty first-level header ("#") to force a new page; the empty header will be removed from the table of contents. The Pandoc filter will now abort if it sees %newpage%.
• ePub is now ePub v3, not ePub v2.
• Added build logic to allow overriding HTML and/or ePub styling by creating book/html.css and/or book/epub.css.
• Ensured that generated ePub passes
  EpubCheck with no errors.
• Fixed ePub CSS file to be proper CSS.
• Removed stray styling in ePub CSS that was preventing the correct paragraph style.
• Corrected generation of ePub metadata so that a lack of a book identifier doesn't generate an empty <dc:identifier> element. Necessary to pass EpubCheck validation.
• Created a new sample cover image, at a higher resolution. Modified LaTeX logic to scale it down properly for PDF.
• Cleaned build file up a bit.

Version 0.3.0

• Added support for generating a bibiography (references) section, appendices,
  a foreward, a preface, and a glossary. All are optional.
• Reworked how the pandoc filter handles token substitution.
• Moved metadata to a Pandoc-style metadata file.
• Added more substitution tokens.
• Added version target to build.
• Added an upgrade script, to help with upgrading to new versions.
• Added code to insert inline cover image in HTML version of the book.

Version 0.2.0

• Enhanced HTML output considerably, based on the "GitHub Pandoc CSS" in this gist.
• Reorganized files so the top directory isn't so cluttered.
• Changed the HTML build process to inline the CSS, to make the HTML output truly standalone.