From 2002186eeddb5bbbab14ad625956734517baa9ed Mon Sep 17 00:00:00 2001
From: fredck <f.knabben@cksource.com>
Date: Wed, 18 Dec 2019 18:32:01 +0100
Subject: [PATCH 1/5] Added test for code blocks with triple/quadruple ticks
 inside.

---
 tests/gfmdataprocessor/code.js | 36 ++++++++++++++++++++++++++++++++++
 1 file changed, 36 insertions(+)

diff --git a/tests/gfmdataprocessor/code.js b/tests/gfmdataprocessor/code.js
index b3dd7df..9af7bc9 100644
--- a/tests/gfmdataprocessor/code.js
+++ b/tests/gfmdataprocessor/code.js
@@ -276,5 +276,41 @@ describe( 'GFMDataProcessor', () => {
 				'`code `` code ``` `'
 			);
 		} );
+
+		it( 'should handle triple ticks inside code', () => {
+			testDataProcessor(
+				'````\n' +
+				'```\n' +
+				'Code\n' +
+				'```\n' +
+				'````',
+
+				'<pre><code>' +
+				'```\n' +
+				'Code\n' +
+				'```' +
+				'</code></pre>'
+			);
+		} );
+
+		it( 'should handle triple and quatruple ticks inside code', () => {
+			testDataProcessor(
+				'`````\n' +
+				'````\n' +
+				'```\n' +
+				'Code\n' +
+				'```\n' +
+				'````\n' +
+				'`````',
+
+				'<pre><code>' +
+				'````\n' +
+				'```\n' +
+				'Code\n' +
+				'```\n' +
+				'````' +
+				'</code></pre>'
+			);
+		} );
 	} );
 } );

From 72174d9aac1733b0922cbec8269aac64c95660e3 Mon Sep 17 00:00:00 2001
From: fredck <f.knabben@cksource.com>
Date: Tue, 25 Feb 2020 16:09:20 +0100
Subject: [PATCH 2/5] Fixed the html->markdown conversion of code blocks with
 multiple ticks inside.

---
 src/html2markdown/html2markdown.js | 39 +++++++++++++++++++++++++++++-
 1 file changed, 38 insertions(+), 1 deletion(-)

diff --git a/src/html2markdown/html2markdown.js b/src/html2markdown/html2markdown.js
index 88d4afd..2bffb93 100644
--- a/src/html2markdown/html2markdown.js
+++ b/src/html2markdown/html2markdown.js
@@ -27,7 +27,8 @@ const turndownService = new TurndownService( {
 
 turndownService.use( [
 	gfm,
-	todoList
+	todoList,
+	fencedCodeBlock
 ] );
 
 export default function html2markdown( html ) {
@@ -47,3 +48,39 @@ function todoList( turndownService ) {
 		}
 	} );
 }
+
+// This one fixes https://github.com/domchristie/turndown/issues/300. It's based on:
+// https://github.com/domchristie/turndown/blob/cae7098f97bcf14118a916e13e807536f432f3ac/src/commonmark-rules.js#L101-L121
+function fencedCodeBlock( turndownService ) {
+	turndownService.addRule( 'taskListItems', {
+		filter( node, options ) {
+			return (
+				options.codeBlockStyle === 'fenced' &&
+				node.nodeName === 'PRE' &&
+				node.firstChild &&
+				node.firstChild.nodeName === 'CODE'
+			);
+		},
+
+		replacement( content, node ) {
+			const className = node.firstChild.className || '';
+			const language = ( className.match( /language-(\S+)/ ) || [ null, '' ] )[ 1 ];
+			const code = node.firstChild.textContent;
+			let fenceSize = 3;
+
+			for ( const match of code.matchAll( /^`{3,}/gm ) ) {
+				if ( match[ 0 ].length >= fenceSize ) {
+					fenceSize = match[ 0 ].length + 1;
+				}
+			}
+
+			const fence = '`'.repeat( fenceSize );
+
+			return (
+				'\n\n' + fence + language + '\n' +
+				code +
+				'\n' + fence + '\n\n'
+			);
+		}
+	} );
+}

From 19446ee6ed2a9ffc9985316d36722f96e8093ad0 Mon Sep 17 00:00:00 2001
From: fredck <f.knabben@cksource.com>
Date: Mon, 9 Mar 2020 13:10:12 +0100
Subject: [PATCH 3/5] Fixed the html->markdown conversion to avoid escaping
 urls in text nodes.

---
 src/html2markdown/html2markdown.js | 39 ++++++++++++++++++++++++---
 tests/gfmdataprocessor/text.js     | 43 ++++++++++++++++++++++++++++++
 2 files changed, 78 insertions(+), 4 deletions(-)
 create mode 100644 tests/gfmdataprocessor/text.js

diff --git a/src/html2markdown/html2markdown.js b/src/html2markdown/html2markdown.js
index 2bffb93..51bf635 100644
--- a/src/html2markdown/html2markdown.js
+++ b/src/html2markdown/html2markdown.js
@@ -10,12 +10,43 @@ import { gfm } from 'turndown-plugin-gfm';
 {
 	const originalEscape = TurndownService.prototype.escape;
 	TurndownService.prototype.escape = function( string ) {
-		string = originalEscape( string );
+		// Urls should not be escaped. Our strategy is using a regex to find them and escape everything
+		// which is out of the matches parts.
 
-		// Escape "<".
-		string = string.replace( /</g, '\\<' );
+		// eslint-disable-next-line max-len
+		const regex = /\b(?:https?:\/\/|www\.)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()[\]{};:'".,<>?«»“”‘’])/g;
 
-		return string;
+		let escaped = '';
+		let lastIndex = 0;
+		let m;
+		do {
+			m = regex.exec( string );
+
+			// The substring should to to the matched index or, if nothing found, the end of the string.
+			const index = m ? m.index : string.length;
+
+			// Append the substring between the last match and the current one (if anything).
+			if ( index > lastIndex ) {
+				escaped += escape( string.substring( lastIndex, index ) );
+			}
+
+			// Append the match itself now, if anything.
+			m && ( escaped += m[ 0 ] );
+
+			lastIndex = regex.lastIndex;
+		}
+		while ( m );
+
+		return escaped;
+
+		function escape( string ) {
+			string = originalEscape( string );
+
+			// Escape "<".
+			string = string.replace( /</g, '\\<' );
+
+			return string;
+		}
 	};
 }
 
diff --git a/tests/gfmdataprocessor/text.js b/tests/gfmdataprocessor/text.js
new file mode 100644
index 0000000..f8770fd
--- /dev/null
+++ b/tests/gfmdataprocessor/text.js
@@ -0,0 +1,43 @@
+/**
+ * @license Copyright (c) 2003-2020, CKSource - Frederico Knabben. All rights reserved.
+ * For licensing, see LICENSE.md or https://ckeditor.com/legal/ckeditor-oss-license
+ */
+
+import { testDataProcessor } from '../_utils/utils';
+
+describe( 'GFMDataProcessor', () => {
+	describe( 'text', () => {
+		describe( 'urls', () => {
+			it( 'should not escape urls', () => {
+				testDataProcessor(
+					'escape\\_this https://test.com/do_[not]-escape escape\\_this',
+					'<p>escape_this https://test.com/do_[not]-escape escape_this</p>'
+				);
+			} );
+
+			it( 'should not escape urls (at start)', () => {
+				testDataProcessor(
+					'https://test.com/do_[not]-escape escape\\_this',
+					'<p>https://test.com/do_[not]-escape escape_this</p>'
+				);
+			} );
+
+			it( 'should not escape urls (at end)', () => {
+				testDataProcessor(
+					'escape\\_this https://test.com/do_[not]-escape',
+					'<p>escape_this https://test.com/do_[not]-escape</p>'
+				);
+			} );
+
+			[
+				'https://test.com/do_[not]-escape',
+				'http://test.com/do_[not]-escape',
+				'www.test.com/do_[not]-escape'
+			].forEach( url => {
+				it( `should not escape urls (${ url })`, () => {
+					testDataProcessor( url, `<p>${ url }</p>` );
+				} );
+			} );
+		} );
+	} );
+} );

From 88004bd11b98ff48993f896a70ed99b0258360f5 Mon Sep 17 00:00:00 2001
From: fredck <f.knabben@cksource.com>
Date: Fri, 13 Mar 2020 17:09:44 +0100
Subject: [PATCH 4/5] Upgraded to Turndown v6.0.0.

---
 package.json | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/package.json b/package.json
index bf502d4..1b9dc46 100644
--- a/package.json
+++ b/package.json
@@ -12,7 +12,7 @@
   "dependencies": {
     "@ckeditor/ckeditor5-engine": "^17.0.0",
     "marked": "^0.7.0",
-    "turndown": "^5.0.3",
+    "turndown": "^6.0.0",
     "turndown-plugin-gfm": "^1.0.2"
   },
   "devDependencies": {

From 43c79726ad47875d6c2bfa4f5af0bf3275045bac Mon Sep 17 00:00:00 2001
From: fredck <f.knabben@cksource.com>
Date: Fri, 13 Mar 2020 17:10:54 +0100
Subject: [PATCH 5/5] Revert "Fixed the html->markdown conversion of code
 blocks with multiple ticks inside.". Turndown v6.0.0 upgrade follow-up.

This reverts commit 72174d9a.
---
 src/html2markdown/html2markdown.js | 39 +-----------------------------
 1 file changed, 1 insertion(+), 38 deletions(-)

diff --git a/src/html2markdown/html2markdown.js b/src/html2markdown/html2markdown.js
index 2bffb93..88d4afd 100644
--- a/src/html2markdown/html2markdown.js
+++ b/src/html2markdown/html2markdown.js
@@ -27,8 +27,7 @@ const turndownService = new TurndownService( {
 
 turndownService.use( [
 	gfm,
-	todoList,
-	fencedCodeBlock
+	todoList
 ] );
 
 export default function html2markdown( html ) {
@@ -48,39 +47,3 @@ function todoList( turndownService ) {
 		}
 	} );
 }
-
-// This one fixes https://github.com/domchristie/turndown/issues/300. It's based on:
-// https://github.com/domchristie/turndown/blob/cae7098f97bcf14118a916e13e807536f432f3ac/src/commonmark-rules.js#L101-L121
-function fencedCodeBlock( turndownService ) {
-	turndownService.addRule( 'taskListItems', {
-		filter( node, options ) {
-			return (
-				options.codeBlockStyle === 'fenced' &&
-				node.nodeName === 'PRE' &&
-				node.firstChild &&
-				node.firstChild.nodeName === 'CODE'
-			);
-		},
-
-		replacement( content, node ) {
-			const className = node.firstChild.className || '';
-			const language = ( className.match( /language-(\S+)/ ) || [ null, '' ] )[ 1 ];
-			const code = node.firstChild.textContent;
-			let fenceSize = 3;
-
-			for ( const match of code.matchAll( /^`{3,}/gm ) ) {
-				if ( match[ 0 ].length >= fenceSize ) {
-					fenceSize = match[ 0 ].length + 1;
-				}
-			}
-
-			const fence = '`'.repeat( fenceSize );
-
-			return (
-				'\n\n' + fence + language + '\n' +
-				code +
-				'\n' + fence + '\n\n'
-			);
-		}
-	} );
-}