MarkAppend is a fast, single-file Markdown parser designed for simplicity and performance. Unlike most Markdown parsers that generate HTML strings, MarkAppend modifies the DOM directly, making it ideal for dynamic web applications that need the best performance.
MarkAppend supports all the essential CommonMark features, skipping only the more esoteric rules that most users aren't aware of โ mainly to keep the implementation simple(r).
appendMarkdown
is the only one exported function in the library. Its signature
looks like this:
export function appendMarkdown(input: string, root: Element)
The input
string parameter contains the markdown that is rendered under the
root
element.
MarkAppend source is written in a literate style, which hopefully makes the code easier to decipher. You can explore the documentation generated from the source code here. The implementation uses regular expressions heavily and is very imperative in style. Special attention is paid to minimize string copying and unnecessary memory allocations.
As the Markdown syntax is inherently ill-defined and ambigious, parsing it inevitably becomes a convoluted process. Lot of custom code is needed to handle all the special cases. So, don't expect to find a beautiful and easy-to-follow implementation ๐ฉ
It's possible to extend the parser with new inline elements, by adding a new Parser. The parser constist of a regular expression and a matcher function that is called when the expresion is found in the input.
Below is an example, how to add support for equations using the AsciiMath2ML library.
import * as ma from 'markappend'
import * as am from 'asciimath2ml'
pr.addInlineParser({
regexp: /(?<![\\$])(?<eqdelim>\$\$?)(?!\$)(?<eq>.+)(?<![\\$])\k<eqdelim>(?!\$)/.source,
matched: (state, match) => {
let { eqdelim, eq } = match.groups!
let html = am.asciiToMathML(eq, eqdelim.length == 1)
pr.appendHtml(state, html)
}
})
You can test MarkAppend with the live editor below.