4 minute read Published:

Build Status

Mmark is a powerful markdown processor written in Go, geared towards writing IETF documents. It is, however, also suited for writing complete books and other technical documentation, like the Learning Go book (mmark source, and I-D text output).

Also see this repository on how to write RFC using Markdown.

It provides an advanced markdown dialect that processes file(s) to produce internet-drafts in XML RFC 7991 format. Mmark can produce xml2rfc (aforementioned RFC 7991), RFC 7749 (xml2rfc version 2 - now deprecated), HTML5 output, markdown and manual pages.

Example RFCs in Mmark format can be found in the Github repository.

Mmark uses gomarkdown which is a fork of blackfriday. See its README.md for more documentation.

If you like Go and parsing text, drop me (miek@miek.nl) a line if you want to be part of the Mmarkdown Github org, and help develop Mmark!


Mmark’s syntax and the extra features compared to plain Markdown are detailed in syntax.md.

Mmark adds the following syntax elements to gomarkdown/markdown:


You can download a binary or optionally build mmark your self. You’ll need a working Go environment, then check out the code and:

% go get && go build
% ./mmark -version

To output XML2RFC v3 xml just give it a markdown file and:

% ./mmark rfc/3514.md

Making a draft in text form (v3 output)

% ./mmark rfc/3514.md > x.xml
% xml2rfc --v3 --text x.xml

Outputting HTML5 is done with the -html switch. Outputting markdown is done with the -markdown switch (optionally you can use -width to set the text width).

Example RFC

The rfc/ directory contains a couple of example RFCs that can be build via the v2 or v3 tool chain. The build the text files, just run:

cd rfc
make txt

Official RFCs are in rfc/orig (so you can compare the text output from mmark).

Using Mmark as a library

By default Mmark gives you a binary you can run, if you want to include the parser and renderers in your own code you’ll have to lift some of it out of mmark.go.

Create a parser with the correct options and flags. The that init is used to track file includes. In this snippet we set if to fileName which is the file we’re currently reading. If reading from standard input, this can be set to "".

p := parser.NewWithExtensions(mparser.Extensions)
init := mparser.NewInitial(fileName)
documentTitle := "" // hack to get document title from TOML title block and then set it here.
p.Opts = parser.Options{
    ParserHook: func(data []byte) (ast.Node, []byte, int) {
        node, data, consumed := mparser.Hook(data)
        if t, ok := node.(*mast.Title); ok {
            if !t.IsTriggerDash() {
                documentTitle = t.TitleData.Title
        return node, data, consumed
    ReadIncludeFn: init.ReadInclude,
    Flags:         parserFlags,

Then parser the document (d is a []byte containing the document text):

doc := markdown.Parse(d, p)

After this doc is ready to be rendered. Create a renderer, with a bunch of options.

opts := html.RendererOptions{
    Comments:       [][]byte{[]byte("//"), []byte("#")}, // used for callouts.
	RenderNodeHook: mhtml.RenderHook,
	Flags:          html.CommonFlags | html.FootnoteNoHRTag | html.FootnoteReturnLinks| html.CompletePage,
	Generator:      `  <meta name="GENERATOR" content="github.com/mmarkdown/mmark Mmark Markdown Processor - mmark.miek.nl`,
opts.Title = documentTitle // hack to add-in discovered title

renderer := html.NewRenderer(opts)

Next we we only need to generate the HTML: x := markdown.Render(doc, renderer). Now x contains a []byte with the HTML.

Also See

Kramdown-rfc2629 is another tool to process markdown and output XML2RFC XML.