gutenberg-markdown-converter.diff

diff --git a/blocks/api/raw-handling/markdown-converter.js b/blocks/api/raw-handling/markdown-converter.js
new file mode 100644
index 00000000..a2d3fa63
--- /dev/null
+++ b/blocks/api/raw-handling/markdown-converter.js
@@ -0,0 +1,77 @@
+/**
+ * External dependencies
+ */
+import showdown from 'showdown';
+
+export default class MarkdownConverter {
+	makeHtml( html ) {
+		// Oh, the horror!
+		this.converter = makeConverter( html );
+		return this.converter.makeHtml( html );
+	}
+}
+
+function makeConverter( html ) {
+	let options = {};
+	const tags = collectShortcodes( html );
+	if ( tags.length ) {
+		const regex = tagsToRegex( collectShortcodes( html ) );
+		showdown.extension( 'ignoreShortcodes', ignoreShortcodes( regex ) );
+		options.extensions = [ 'ignoreShortcodes' ];
+	}
+	const converter = new showdown.Converter( options );
+	converter.setOption( 'noHeaderId', true );
+	converter.setOption( 'tables', true );
+	converter.setOption( 'literalMidWordUnderscores', true );
+	return converter;
+}
+
+function collectShortcodes( html ) {
+	let match;
+	const tags = [];
+	const re = /\[\/([a-z][a-z0-9_-]*)\]/g;
+	while ( ( match = re.exec( html ) ) !== null ) {
+		tags.push( match[ 1 ] );
+	}
+	return tags;
+}
+
+function tagsToRegex( tags ) {
+	return new RegExp(
+		'(' +
+		// TODO regex-escape tag
+		tags.map( tag => (
+			`\\[${tag}\\](.*)\\[\\/${tag}\\]`
+		) ).join( '|' ) +
+		')',
+		'g'
+	);
+}
+
+function ignoreShortcodes( regex ) {
+	return function() {
+		let matches = [];
+		return [
+			{ 
+				type: 'lang',
+				regex,
+				replace( string ) { 
+					matches.push( string );
+					const n = matches.length - 1;
+					return '%PLACEHOLDER' + n + '%';
+				},
+			},
+			{
+				type: 'output',
+				filter(text) {
+					for (let i=0; i< matches.length; ++i) {
+						const pat = '%PLACEHOLDER' + i + '% *';
+						text = text.replace(new RegExp(pat, 'g'), matches[i]);
+					}
+					matches = [];
+					return text;
+				},
+			}
+		];
+	};
+}
diff --git a/blocks/api/raw-handling/test/markdown-converter.js b/blocks/api/raw-handling/test/markdown-converter.js
new file mode 100644
index 00000000..23882eff
--- /dev/null
+++ b/blocks/api/raw-handling/test/markdown-converter.js
@@ -0,0 +1,38 @@
+/**
+ * External dependencies
+ */
+import { equal } from 'assert';
+
+/**
+ * Internal dependencies
+ */
+import MarkdownConverter from '../markdown-converter';
+
+describe( 'MarkdownConverter', () => {
+	describe( 'when using shortcode extension', () => {
+		const converter = new MarkdownConverter();
+
+		it( 'should ignore formatting inside blacklisted shortcodes', () => {
+			const input1 = '[my_shortcode]hello, _there_.[/my_shortcode]';
+			const output1 = '<p>[my_shortcode]hello, _there_.[/my_shortcode]</p>';
+			equal( converter.makeHtml( input1 ), output1 );
+
+			const input2 = '[other_shortcode]hello, _there_.[/other_shortcode]';
+			const output2 = '<p>[other_shortcode]hello, _there_.[/other_shortcode]</p>';
+			equal( converter.makeHtml( input2 ), output2 );
+		} );
+
+		it( 'should do the same for multi-occurrence strings', () => {
+			const input = '[my_shortcode]hello, _there_.[/my_shortcode] [other_shortcode]hello, _there_.[/other_shortcode]';
+
+			const output = '<p>[my_shortcode]hello, _there_.[/my_shortcode][other_shortcode]hello, _there_.[/other_shortcode]</p>';
+			equal( converter.makeHtml( input ), output );
+		} );
+
+		it( 'should ignore content around unclosed shortcode tags', () => {
+			const input = '**oh**, [unknown]hello, _there_.';
+			const output = '<p><strong>oh</strong>, [unknown]hello, <em>there</em>.</p>';
+			equal( converter.makeHtml( input ), output );
+		} );
+	} );
+} );