Showdown is a Javascript Markdown to HTML converter, based on the original works by John Gruber. Showdown can be used client side (in the browser) or server side (with NodeJs).
In theory, Showdown will work in any browser that supports ECMA 262 3rd Edition (JavaScript 1.5). The converter itself might even work in things that aren't web browsers, like Acrobat. No promises.
A suite of tests is available which require node.js. Once node is installed, run the following command from the project root to install the development dependencies:
npm install --dev
Once installed the tests can be run from the project root using:
npm test
New test cases can easily be added. Create a markdown file (ending in `.md`) which contains the markdown to test. Create a `.html` file of the exact same name. It will automatically be tested when the tests are executed with `mocha`.
In most cases, Showdown's output is identical to that of Perl Markdown v1.0.2b7. What follows is a list of all known deviations. Please file an issue if you find more.
* This release uses the HTML parser from Markdown 1.0.2b2,
which means it fails `Inline HTML (Advanced).text` from
the Markdown test suite:
<div>
<div>
unindented == broken
</div>
</div>
* Showdown doesn't support the markdown="1" attribute:
<divmarkdown="1">
Markdown does *not* work in here.
</div>
This is half laziness on my part and half stubbornness.
Markdown is smart enough to process the contents of span-
level tags without screwing things up; shouldn't it be
able to do the same inside block elements? Let's find a
way to make markdown="1" the default.
* You can only nest square brackets in link titles to a
depth of two levels:
[[fine]](http://www.attacklab.net/)
[[[broken]]](http://www.attacklab.net/)
If you need more, you can escape them with backslashes.
* When sublists have paragraphs, Showdown produces equivalent
HTML with a slightly different arrangement of newlines:
+ item
- subitem
The HTML has a superfluous newline before this
paragraph.
- subitem
The HTML here is unchanged.
- subitem
The HTML is missing a newline after this
list subitem.
* Markdown.pl creates empty title attributes for
inline-style images:
Here's an empty title on an inline-style
.
I tried to replicate this to clean up my diffs during
testing, but I went too far: now Showdown also makes
empty titles for reference-style images:
Showdown makes an empty title for
reference-style ![images][] too.
[images]: http://w3.org/Icons/valid-xhtml10
* With crazy input, Markdown will mistakenly put
`<strong>` or `<em>` tags in URLs:
<ahref="<*Markdown adds em tags in here*>">
improbable URL
</a>
Showdown won't. But still, don't do that.
## Creating Markdown Extensions
A showdown extension is simply a function which returns an array of extensions. Each single extension can be one of two types:
* Language Extension -- Language extensions are ones that that add new markdown syntax to showdown. For example, say you wanted `^^youtube http://www.youtube.com/watch?v=oHg5SJYRHA0` to automatically render as an embedded YouTube video, that would be a language extension.
* Output Modifiers -- After showdown has run, and generated HTML, an output modifier would change that HTML. For example, say you wanted to change `<div class="header">` to be `<header>`, that would be an output modifier.
Each extension can provide two combinations of interfaces for showdown.
### Regex/Replace
Regex/replace style extensions are very similar to Javascript's `string.replace` function. Two properties are given, `regex` and `replace`. `regex` is a string and `replace` can be either a string or a function. If `replace` is a string, it can use the `$1` syntax for group substitution, exactly as if it were making use of `string.replace` (internally it does this actually); The value of `regex` is assumed to be a global replacement.
**Example:**
```js
var demo = function(converter) {
return [
// Replace escaped @ symbols
{ type: 'lang', regex: '\\@', replace: '@' }
];
}
```
### Filter
Alternately, if you'd just like to do everything yourself, you can specify a filter which is a callback with a single input parameter, text (the current source text within the showdown engine).
**Example:**
```js
var demo = function(converter) {
return [
// Replace escaped @ symbols
{ type: 'lang', filter: function(text) {
return text.replace(/\\@/g, '@');
}}
];
}
```
### Implementation Concerns
One bit which should be taken into account is maintaining both client-side and server-side compatibility. This can be achieved with a few lines of boilerplate code. First, to prevent polluting the global scope for client-side code, the extension definition should be wrapped in a self-executing function.
```js
(function(){
// Your extension here
}());
```
Second, client-side extensions should add a property onto `Showdown.extensions` which matches the name of the file. As an example, a file named `demo.js` should then add `Showdown.extensions.demo`. Server-side extensions can simply export themselves.
