Convert DLR documentation from Word to Markdown for inclusion on the DLR repo wiki.
Some background at IronLanguages/dlr#246.
This project carries out the following (everything takes place in the Filter project, unless noted otherwise):
- Identify and mark code blocks in such a way that they can be recognized by Pandoc (
Driverproject) - Generate the per-Word-document TOC into
_sidebar.md - Split each Word document into multiple Markdown files, based on the 1st-level header. (This involves multiple passes over the document, each time returning to Pandoc only the relevant section)
- Document headings are auto-numbered; Pandoc doesn't recreate this numbering in Markdown. Recreate them in the text, but not in the generated IDs. This entails using embedded HTML to explicitly write the IDs
- Replace HTML markup for images with Markdown, by removing positioning information
- Replace HTML markup for underlines with Markdown bold
- Trim excess spaces from beginning of lines in code blocks
- Reduce bullet list spacing from the Pandoc-generated 3 spaces to 1 space (
Driverproject)