Skip to content

Node.js bindings to GitHub's GFM-enhanced fork of cmark, the CommonMark reference implementation in C

Notifications You must be signed in to change notification settings

0x777/node-cmark-gfm

 
 

Repository files navigation

cmark-gfm

cmark-gfm is a Node.js wrapper around libcmark-gfm, GitHub's GFM-enhanced fork of cmark, the reference implementation of CommonMark in C by John MacFarlane.

Installation

cmark-gfm is distributed via the npm registry, and can be installed with npm and similar tools:

npm install cmark-gfm

API

renderHtml(markdown[, options][, callback])

Converts a Markdown string to HTML asynchronously. If you do not provide a callback, renderHtml will return a Promise that will resolve to the resulting HTML. If options is omitted, default options will be used.

  • markdown - a string of Markdown to convert to HTML
  • options - a hash of options (see Options, below)
  • callback - a Node.js-style callback to call with the resulting HTML once the Markdown has been rendered
    • error - any error that occurred
    • html - the resulting HTML

Example:

const cmark = require('cmark-gfm')

const options = { ... }

// Promise-style
cmark.renderHtml('# Hello there!', options)
  .then(html => console.log(html))
  .catch(error => console.error(error))

// Callback-style
cmark.renderHtml('# Hello there!', options, (error, html) => {
  if (error) {
    console.error(error)
    return
  }

  console.log(html)
})

html = renderHtmlSync(markdown[, options])

Converts a Markdown string to HTML synchronously. If options is omitted, default options will be used.

  • markdown - a string of Markdown to convert to HTML
  • options - a hash of options (see Options, below)

Example:

const cmark = require('cmark-gfm')

const options = { ... }

try {
  const html = cmark.renderHtmlSync('# Hello there!', options)
  console.log(html)
} catch (error) {
  console.error(error)
}

createStreamingParser([options])

Creates a stream with a writable end that accepts Markdown and a readable end that produces HTML. The parser ingests Markdown and converts it to HTML asynchronously. If options is omitted, default options will be used.

  • options - a hash of options (see Options, below)

Example:

const cmark = require('cmark-gfm')
const fs = require('fs')

const options = { ... }

fs.createReadStream('./input.md')
  .pipe(cmark.createStreamingParser(options))
  .pipe(fs.createWriteStream('./output.html'))

Options

cmark-gfm supports a variety of options that it passes to the underlying libcmark-gfm library. You can enable most options by setting its name to the value of true in your options object; to enable an extension, add its name as a key to an extensions object with a value of true. For example, to enable the smart and footnotes options and the strikethrough extension:

const cmark = require('cmark-gfm')

const options = {
  smart: true,
  footnotes: true,
  extensions: {
    strikethrough: true
  }
}

cmark.renderHtml(markdown, options)
  .then(/* ... */)

You can find a summary of all the options in the table below, as well as additional information for some options in the Features section later in this document.

Name Type Description
sourcepos* bool Adds a data-sourcepos attribute to elements that indicate the range of Markdown that resulted in the element
hardbreaks* bool Renders softbreak elements as hard line breaks
nobreaks* bool Renders softbreak elements as spaces
validateUtf8 bool Replaces illegal UTF-8 sequences with U+FFFD
smart bool Replaces straight quotes with curly ones, and turns --- into em dashes and -- into en dashes
githubPreLang* bool Uses GitHub style <pre lang="x"> tags for code blocks
liberalHtmltag bool Allows HTML tags to be parsed as HTML even if they are not well formed (e.g. < div> instead of just <div>)
footnotes* bool Enables footnote parsing and rendering
strikethroughDoubleTilde bool When enabled, the strikethrough extension will only render text as strikethrough if it is surrounded by exactly ~~two tildes~~
fullInfoString* bool Adds additional code block info as an additional attribute on the resulting HTML element
unsafe* bool Allows raw HTML and unsafe links (javascript:, vbscript:, file:, and data: except for image/png, image/gif, image/jpeg, or image/webp mime types). Otherwise, raw HTML is replaced by a placeholder HTML comment, and unsafe links are replaced with empty strings.
extensions* object Which extensions to enable
* more information in the Features and Extensions sections below

libcmark-gfm also exposes several Markdown extensions that you can enable by passing their name as keys to the extensions option (with a value of true). You can find a summary of all the extensions in the table below, as well as additional information in the Extensions section later in this document.

Name Description
"table" Renders tables
"strikethrough" Renders text as strikethrough when surrounded by ~tildes~
"tagfilter" Escapes certain dangerous HTML tags, causing them to have no effect
"autolink" Automatically generates links from URLs
"tasklist" Renders GitHub-style task lists

Features

sourcepos

Enabling the sourcepos option adds a data-sourcepos attribute to block level elements which indicate the row and column positions from the original Markdown source that resulted in that HTML element. For example, given the markdown:

# Hello!

Hi there. This is a Markdown file. You probably knew that already.

```javascript
// Here's a bit of JS code

console.log('hello')
```

And a short list:

* One
* Two
* Three

Rendering with the sourcepos attribute would result in output similar to:

<h1 data-sourcepos="1:1-1:8">Hello!</h1>

<p data-sourcepos="3:1-3:66">Hi there. This is a Markdown file. You probably knew that already.</p>

<pre data-sourcepos="5:1-9:3">
<code class="language-javascript">
// Here's a bit of JS code

console.log('hello')
</code>
</pre>

<p data-sourcepos="11:1-11:17">And a short list:</p>

<ul data-sourcepos="13:1-15:7">
  <li data-sourcepos="13:1-13:5">One</li>
  <li data-sourcepos="14:1-14:5">Two</li>
  <li data-sourcepos="15:1-15:7">Three</li>
</ul>

You can see, for example, that the <pre> tag that represents the code block was generated from the Markdown text between row 5 column 1 and row 9 column 3 (inclusive and 1-indexed).

hardbreaks

Enabling the hardbreaks option renders "softbreak" elements as hard line breaks.

abc
def

would normally result in

<p>abc\ndef</p>

but with hardbreaks enabled, it results in

<p>abc<br>\ndef</p>

nobreaks

Enabling the nobreaks option renders "softbreak" elements as spaces:

abc
def

would normally result in

<p>abc\ndef</p>

but with nobreaks enabled, it results in

<p>abc def</p>

githubPreLang

Normally, code blocks like the following:

```javascript
console.log('hello!')
```

would result in HTML like this:

<pre>
<code class="language-javascript">
console.log('hello!')
</code>
</pre>

With githubPreLang enabled, it would instead result in the following HTML:

<pre lang="javascript">
<code>
console.log('hello!')
</code>
</pre>

footnotes

Enabling the footnotes option turns on the parsing and rendering of footnotes. Here's a simple example:

Here is some text[^1]. Here's a little bit more[^more].

[^1]: You can tell because of all the text
[^more]: Just a little bit

And the resulting HTML:

<p>Here is some text<sup class="footnote-ref"><a href="#fn1" id="fnref1">1</a></sup>. Here's a little bit more<sup class="footnote-ref"><a href="#fn2" id="fnref2">2</a></sup>.</p>

<section class="footnotes">
  <ol>
    <li id="fn1">
      <p>You can tell because of all the text <a href="#fnref1" class="footnote-backref"></a></p>
    </li>
    <li id="fn2">
      <p>Just a little bit <a href="#fnref2" class="footnote-backref"></a></p>
    </li>
  </ol>
</section>

fullInfoString

When the fullInfoString feature is enabled, extra text after the language tag for a code block (which is normally discarded) is included in a data-meta attribute on the resulting HTML element:

```javascript here is some more text
console.log('hi')
```
<pre>
<code class="language-javascript" data-meta="here is some more text">
console.log('hi')
</code>
</pre>

unsafe

By default, libcmark-gfm escapes raw HTML and unsafe links:

![img](data:image/gif;base64,abccdefgh)

<div>hello!</div>

[link](javascript:alert('omg'))

would result in:

<p><img alt="img" src="data:image/gif;base64,abccdefgh"></p>
<!-- raw HTML omitted -->
<p><a href="">link</a></p>

Notice that the div and the javascript:alert link have been sanitized. With the unsafe option enabled, the resulting HTML would be:

<p><img alt="img" src="data:image/gif;base64,abccdefgh"></p>
<div>hello!</div>
<p><a href="javascript:alert('omg')">link</a></p>

If you use the unsafe option, you should be sure to do your own sanitization of the resulting HTML.

Extensions

"table"

The table extension allows the parsing and rendering of tables:

|Header|
|------|
|Hello |
<table>
  <thead>
    <tr>
      <th>Header</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>Hello</td>
    </tr>
  </tbody>
</table>

See the table documentation in the GFM docs for more information.

"strikethrough"

The strikethrough extension renders text surrounded by ~one~ or ~~two~~ tildes as strikethrough using the <del> tag. If the strikethroughDoubleTilde option is enabled, text will only be rendered as strikethrough if it is surrounded by exactly ~~two~~ tildes.

"tagfilter"

Certain HTML tags can be particularly troublesome as they change how HTML is interpreted in a way unique to them. The tagfilter extension escapes these tags by replacing the leading < with &lt;. The following markdown:

<div>What a weird <xmp> tag</div>

would result in this HTML:

<div>What a weird &lt;xmp> tag</div>

See the Disallowed Raw HTML documentation in the GFM docs for more information and the list of filtered tags.

"autolink"

The autolink extension recognizes URLs in text and automatically turns them into links:

Visit us at https://github.com!
<p>Visit us at <a href="https://github.com">https://github.com</a>!</p>

See the autolinks documentation in the GFM docs for more information about how libcmark-gfm scans for links.

"tasklist"

The tasklist extension enables the rendering of GitHub-style task lists.

- [ ] Task 1
- [x] Task 2
<ul>
  <li class="task-list-item"><input type="checkbox" disabled="" /> Task 1</li>
  <li class="task-list-item"><input type="checkbox" checked="" disabled="" /> Task 2</li>
</ul>

See the task list documentation in the GFM docs for more information about how libcmark-gfm processes task lists.

About

Node.js bindings to GitHub's GFM-enhanced fork of cmark, the CommonMark reference implementation in C

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • JavaScript 51.8%
  • C++ 31.1%
  • Python 8.5%
  • TypeScript 5.7%
  • Shell 2.5%
  • C 0.4%