Markdown Link Converter

The Markdown link Converter, which filters markdown pages before the processing of doxygen, converts the links in markdown pages. It is set up as input filter in doxygen, if a markdown file is desired to be in the API documentation it only must have the extension .md and be in the INPUT path.

The Markdown link Converter gives each markdown file a header { #header } which is attached to a title and converts the links to refer to this headers. This conversion happens in 2 passes, which is needed because there can be files with no title.

Usage for manual invocation

    markdownlinkconverter [<cmake-cache-file>] <input-file>

The <input-file> parameter must be an absolute path!

Conventions

• Links starting with @ref, # for anchors and http, https or ftp for extern links wont be touched.
• All other links to markdown or arbitrary source files will be converted.
• All links to folders will be altered to the README.md in the Folder. This feature was introduced to be compatible with github, where you can show the content of a folder in combination with the README.md of the containing folder.
• Anchors wont work in imported markdown pages.

github specialities

• github supports source code fences with syntax highlighting which are not recognised by Doxygen. Thus sh after the fence is removed for Doxygen.

Link Validation

internal links

The link validation works with a simple try to fopen the file, which the link refers to.

external links

Every link starting with http, https or ftp will be written to a file named external-links.txt located in your build folder. With the following syntax:

    <file>|<line> col 0 | <url>

Note: Due to the nature of the Markdown Link Converter the file can only be opened in append mode. So delete it and rerun the html build process (make clean could be needed) to get a list without duplicates.

In the script folder is a script named link-checker. This script can be used to validate the links. Broken links will be printed. False positive not excluded (very rare).

Further improvements (which will be introduced in a later version):

• optimize pdf output (also UTF-8 encoding)
• if title contains –, this should be* also remove other fences doxygen does not understand