Awesome
Comrak
Rust port of github's cmark-gfm
.
Compliant with CommonMark 0.31.2 in default mode.
GFM support synced with release 0.29.0.gfm.13
.
Installation
Specify it as a requirement in Cargo.toml
:
[dependencies]
comrak = "0.29"
Comrak's library supports Rust <span class="msrv">1.62.1</span>+.
CLI
- Anywhere with a Rust toolchain:
cargo install comrak
- Many Unix distributions:
pacman -S comrak
brew install comrak
dnf install comrak
nix run nixpkgs#comrak
You can also find builds I've published in GitHub Releases, but they're limited to machines I have access to at the time of making them! webinstall.dev offers curl | shell
-style installation of the latest of these for your OS.
Usage
<details> <summary>Click to expand the CLI <code>--help</code> output.$ comrak --help
</summary>
A 100% CommonMark-compatible GitHub Flavored Markdown parser and formatter
Usage: comrak [OPTIONS] [FILE]...
Arguments:
[FILE]...
CommonMark file(s) to parse; or standard input if none passed
Options:
-c, --config-file <PATH>
Path to config file containing command-line arguments, or 'none'
[default: /home/runner/.config/comrak/config]
-i, --inplace
To perform an in-place formatting
--hardbreaks
Treat newlines as hard line breaks
--smart
Use smart punctuation
--github-pre-lang
Use GitHub-style <pre lang> for code blocks
--full-info-string
Enable full info strings for code blocks
--gfm
Enable GitHub-flavored markdown extensions: strikethrough, tagfilter, table, autolink, and
tasklist. Also enables --github-pre-lang and --gfm-quirks
--gfm-quirks
Enables GFM-style quirks in output HTML, such as not nesting <strong> tags, which
otherwise breaks CommonMark compatibility
--relaxed-tasklist-character
Enable relaxing which character is allowed in a tasklists
--relaxed-autolinks
Enable relaxing of autolink parsing, allow links to be recognized when in brackets and
allow all url schemes
--default-info-string <INFO>
Default value for fenced code block's info strings if none is given
--unsafe
Allow raw HTML and dangerous URLs
--gemojis
Translate gemojis into UTF-8 characters
--escape
Escape raw HTML instead of clobbering it
--escaped-char-spans
Wrap escaped characters in span tags
-e, --extension <EXTENSION>
Specify extension name(s) to use
Multiple extensions can be delimited with ",", e.g. --extension strikethrough,table
[possible values: strikethrough, tagfilter, table, autolink, tasklist, superscript,
footnotes, description-lists, multiline-block-quotes, math-dollars, math-code,
wikilinks-title-after-pipe, wikilinks-title-before-pipe, underline, spoiler, greentext]
-t, --to <FORMAT>
Specify output format
[default: html]
[possible values: html, xml, commonmark]
-o, --output <FILE>
Write output to FILE instead of stdout
--width <WIDTH>
Specify wrap width (0 = nowrap)
[default: 0]
--header-ids <PREFIX>
Use the Comrak header IDs extension, with the given ID prefix
--front-matter-delimiter <DELIMITER>
Ignore front-matter that starts and ends with the given string
--syntax-highlighting <THEME>
Syntax highlighting for codefence blocks. Choose a theme or 'none' for disabling
[default: base16-ocean.dark]
--list-style <LIST_STYLE>
Specify bullet character for lists (-, +, *) in CommonMark output
[default: dash]
[possible values: dash, plus, star]
--sourcepos
Include source position attribute in HTML and XML output
--experimental-inline-sourcepos
Include inline sourcepos in HTML output, which is known to have issues
--ignore-setext
Ignore setext headers
--ignore-empty-links
Ignore empty links
-h, --help
Print help information (use `-h` for a summary)
-V, --version
Print version information
By default, Comrak will attempt to read command-line options from a config file specified by
--config-file. This behaviour can be disabled by passing --config-file none. It is not an error if
the file does not exist.
</details>
And there's a Rust interface. You can use comrak::markdown_to_html
directly:
use comrak::{markdown_to_html, Options};
assert_eq!(markdown_to_html("Hello, **世界**!", &Options::default()),
"<p>Hello, <strong>世界</strong>!</p>\n");
Or you can parse the input into an AST yourself, manipulate it, and then use your desired formatter:
use comrak::nodes::NodeValue;
use comrak::{format_html, parse_document, Arena, Options};
fn replace_text(document: &str, orig_string: &str, replacement: &str) -> String {
// The returned nodes are created in the supplied Arena, and are bound by its lifetime.
let arena = Arena::new();
// Parse the document into a root `AstNode`
let root = parse_document(&arena, document, &Options::default());
// Iterate over all the descendants of root.
for node in root.descendants() {
if let NodeValue::Text(ref mut text) = node.data.borrow_mut().value {
// If the node is a text node, perform the string replacement.
*text = text.replace(orig_string, replacement);
}
}
let mut html = vec![];
format_html(root, &Options::default(), &mut html).unwrap();
String::from_utf8(html).unwrap()
}
fn main() {
let doc = "This is my input.\n\n1. Also [my](#) input.\n2. Certainly *my* input.\n";
let orig = "my";
let repl = "your";
let html = replace_text(&doc, &orig, &repl);
println!("{}", html);
// Output:
//
// <p>This is your input.</p>
// <ol>
// <li>Also <a href="#">your</a> input.</li>
// <li>Certainly <em>your</em> input.</li>
// </ol>
}
For a slightly more real-world example, see how I generate my GitHub user README from a base document with embedded YAML, which itself has embedded Markdown, or check out some of Comrak's dependents on crates.io or on GitHub.
Security
As with cmark
and cmark-gfm
,
Comrak will scrub raw HTML and potentially dangerous links. This change was introduced in Comrak 0.4.0 in support of a
safe-by-default posture, and later adopted by our contemporaries. :)
To allow these, use the unsafe_
option (or --unsafe
with the command line program). If doing so, we recommend the
use of a sanitisation library like ammonia
configured specific to your needs.
Extensions
Comrak supports the five extensions to CommonMark defined in the GitHub Flavored Markdown Spec:
Comrak additionally supports its own extensions, which are yet to be specced out (PRs welcome!):
- Superscript
- Header IDs
- Footnotes
- Description lists
- Front matter
- Multi-line blockquotes
- Math
- Emoji shortcodes
- Wikilinks
- Underline
- Spoiler text
- "Greentext"
By default none are enabled; they are individually enabled with each parse by setting the appropriate values in the
ExtensionOptions
struct.
Plugins
Fenced code block syntax highlighting
You can provide your own syntax highlighting engine.
Create an implementation of the SyntaxHighlighterAdapter
trait, and then provide an instance of such adapter to
Plugins.render.codefence_syntax_highlighter
. For formatting a Markdown document with plugins, use the
markdown_to_html_with_plugins
function, which accepts your plugins object as a parameter.
See the syntax_highlighter.rs
and syntect.rs
examples for more details.
Syntect
syntect
is a syntax highlighting library for Rust. By default, comrak
offers
a plugin for it. In order to utilize it, create an instance of plugins::syntect::SyntectAdapter
and use it in your
Plugins
option.
Related projects
Comrak's design goal is to model the upstream cmark-gfm
as closely as possible
in terms of code structure. The upside of this is that a change in cmark-gfm
has a very predictable change in Comrak.
Likewise, any bug in cmark-gfm
is likely to be reproduced in Comrak. This could be considered a pro or a con,
depending on your use case.
The downside, of course, is that the code often diverges from idiomatic Rust, especially in the AST's extensive use of RefCell
, and while
contributors have made it as fast as possible, it simply won't be as fast as some other CommonMark parsers
depending on your use-case. Here are some other projects to consider:
- Raph Levien's
pulldown-cmark
. It's very fast, uses a novel parsing algorithm, and doesn't construct an AST (but you can use it to make one if you want).cargo doc
uses this, as do many other projects in the ecosystem. - markdown-rs (1.x) looks worth watching.
- Know of another library? Please open a PR to add it!
As far as I know, Comrak is the only library to implement all of the GitHub Flavored Markdown extensions rigorously.
Benchmarking
You'll need to install hyperfine, and CMake if you want to compare against cmark-gfm
.
If you want to just run the benchmark for the comrak
binary itself, run:
make bench-comrak
This will build Comrak in release mode, and run benchmark on it. You will see the time measurements as reported by hyperfine in the console.
The Makefile
also provides a way to run benchmarks for comrak
current state (with your changes), comrak
main branch, cmark-gfm
, pulldown-cmark
and markdown-it.rs
. You'll need CMake, and ensure submodules are prepared.
make bench-all
This will build and run benchmarks across all, and report the time taken by each as well as relative time.
<!-- XXX: The following isn't really true at the moment, due to https://github.com/kivikakk/comrak/issues/339 --> <!-- Apart from this, CI is also setup for running benchmarks when a pull request is first opened. It will add a comment with the results on the pull request in a tabular format comparing the 5 versions. After that you can manually trigger this CI by commenting `/run-bench` on the PR, this will update the existing comment with new results. Note benchmarks won't be automatically run on each push. -->Contributing
Contributions are highly encouraged; if you'd like to assist, consider checking out the good first issue
label! I'm happy to help provide direction and guidance throughout, even if (especially if!) you're new to Rust or open source.
Where possible I practice Optimistic Merging as described by Peter Hintjens. Please keep the code of conduct in mind too.
Thank you to Comrak's many contributors for PRs and issues opened!
Code Contributors
Financial Contributors
Become a financial contributor and help sustain Comrak's development. I'm self-employed --- open-source software relies on the collective.
Contact
Asherah Connor <ashe kivikakk ee>
Legal
Copyright (c) 2017–2024, Asherah Connor and Comrak contributors. Licensed under the 2-Clause BSD License.
cmark
itself is is copyright (c) 2014, John MacFarlane.
See COPYING for all the details.