support for cdata elements like <script> and <xmp>

tmont · Jun 6, 2012 · a83f8a5 · a83f8a5
1 parent 00ad88c
commit a83f8a5
Show file tree

Hide file tree

Showing 5 changed files with 103 additions and 5 deletions.
diff --git a/src/parser.js b/src/parser.js
@@ -58,6 +58,17 @@ function parseOpenElement(context) {
 	var name = context.readRegex(nameRegex);
 	context.callbacks.openElement(name);
 	readAttributes(context, false);
+
+	if (name !== 'script' && name !== 'xmp') {
+		return;
+	}
+
+	//just read until the closing tags for elements that allow cdata
+	var regex = new RegExp('^([\\s\\S]*?)(?:$|</' + name + '>)', 'i');
+	var match = regex.exec(context.substring);
+	context.read(match[0].length);
+	context.callbacks.cdata(match[1]);
+	context.callbacks.closeElement(name);
 }
 
 function parseEndElement(context) {

diff --git a/tests/cdata-tests.js b/tests/cdata-tests.js
@@ -85,4 +85,53 @@ describe('CDATA', function() {
 		cdataCount.should.equal(1);
 		textCount.should.equal(1);
 	});
+
+	it('script tags do not require CDATA', function() {
+		var closeCount = 0, cdataCount = 0, openCount = 0;
+		helpers.parseString('<script><foo></script>', {
+			openElement: function(name) {
+				name.should.equal('script');
+				openCount++;
+			},
+
+			closeElement: function(name) {
+				name.should.equal('script');
+				closeCount++;
+			},
+
+			cdata: function(value) {
+				value.should.equal('<foo>');
+				cdataCount++;
+			}
+		});
+
+		cdataCount.should.equal(1);
+		closeCount.should.equal(1);
+		openCount.should.equal(1);
+	});
+
+	it('xmp tags do not require CDATA', function() {
+		var closeCount = 0, cdataCount = 0, openCount = 0;
+		helpers.parseString('<xmp><foo></xmp>', {
+			openElement: function(name) {
+				name.should.equal('xmp');
+				openCount++;
+			},
+
+			closeElement: function(name) {
+				name.should.equal('xmp');
+				closeCount++;
+			},
+
+			cdata: function(value) {
+				value.should.equal('<foo>');
+				cdataCount++;
+			}
+		});
+
+		cdataCount.should.equal(1);
+		closeCount.should.equal(1);
+		openCount.should.equal(1);
+	});
+
 });
diff --git a/tests/files/good-expected.html b/tests/files/good-expected.html
@@ -3,7 +3,7 @@
 <html>
 	<head>
 		<title>tommy montgomery </title>
-		<meta http-equiv="content-type" content="text/html;charset=utf-8"/>
+		<meta foo="bar" http-equiv="content-type" content="text/html;charset=utf-8"/>
 		<link rel="stylesheet" href="/media/css/global.css"/>
 		<link rel="shortcut icon" type="image/png" href="/favicon.ico"/>
 	</head>
@@ -19,6 +19,14 @@ <h1><a href="/"><span>t</span>ommy <span>mont</span>gomery</a></h1>
 			</div>
 		</div>
 
-		<script src="/media/js/d3.min.js"></script>
+		<!-- this is a comment -->
+		<footer>
+			<![CDATA[ I don't need to escape <script> ]]>
+		</footer>
+
+		<script type="text/javascript">
+			var foo = '<foo>';
+			document.write(foo);
+		</script>
 	</body>
 </html>
diff --git a/tests/files/good.html b/tests/files/good.html
@@ -2,9 +2,12 @@
 <html>
 	<head>
 		<title>tommy montgomery </title>
-		<meta http-equiv="content-type" content="text/html;charset=utf-8"/>
+		<meta foo=bar http-equiv="content-type" content="text/html;charset=utf-8"/>
 		<link rel="stylesheet" href="/media/css/global.css"/>
-		<link rel="shortcut icon" type="image/png" href="/favicon.ico"/>
+		<link
+			rel="shortcut icon"
+			type="image/png"
+			href="/favicon.ico"/>
 	</head>
 
 	<body>
@@ -18,6 +21,14 @@ <h1><a href="/"><span>t</span>ommy <span>mont</span>gomery</a></h1>
 			</div>
 		</div>
 
-		<script src="/media/js/d3.min.js"></script>
+		<!-- this is a comment -->
+		<footer>
+			<![CDATA[ I don't need to escape <script> ]]>
+		</footer>
+
+		<script type="text/javascript">
+			var foo = '<foo>';
+			document.write(foo);
+		</script>
 	</body>
 </html>
diff --git a/tests/integration-tests.js b/tests/integration-tests.js
@@ -4,13 +4,17 @@ var helpers = require('./helpers');
 describe('Integration', function() {
 	it('real life HTML document', function(done) {
 		var content = '';
+		var inScriptTag = false;
 		helpers.parser.parseFile(__dirname + '/files/good.html', 'utf8', {
 			docType: function(value) {
 				content += '<!doctype ' + value + '>\n';
 			},
 
 			openElement: function(name) {
 				content += '<' + name;
+				if (name === 'script') {
+					inScriptTag = true;
+				}
 			},
 
 			closeOpenedElement: function(token) {
@@ -19,6 +23,9 @@ describe('Integration', function() {
 
 			closeElement: function(name) {
 				content += '</' + name + '>';
+				if (name === 'script') {
+					inScriptTag = false;
+				}
 			},
 
 			attribute: function(name, value) {
@@ -27,6 +34,18 @@ describe('Integration', function() {
 
 			text: function(value) {
 				content += value;
+			},
+
+			comment: function(value) {
+				content += '<!--' + value + '-->';
+			},
+
+			cdata: function(value) {
+				if (inScriptTag) {
+					content += value;
+				} else {
+					content += '<![CDATA[' + value + ']]>';
+				}
 			}
 		}, function(err) {
 			should.not.exist(err);