Ensure that the XML prolog doesn't make it into the DOM

rubys · rubys · commit 2edd4504adf2 · 2007-06-27T17:11:23.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40834
diff --git a/src/html5lib/liberalxmlparser.py b/src/html5lib/liberalxmlparser.py
@@ -69,6 +69,7 @@ class XHTMLParser(XMLParser):
 
     def __init__(self, *args, **kwargs):
         html5parser.HTMLParser.__init__(self, *args, **kwargs)
+        self.phases["initial"] = XmlInitialPhase(self, self.tree)
         self.phases["rootElement"] = XhmlRootPhase(self, self.tree)
 
     def normalizeToken(self, token):
@@ -101,7 +102,19 @@ def insertHtmlElement(self):
         self.tree.document.appendChild(element)
         self.parser.phase = self.parser.phases["beforeHead"]
 
+class XmlInitialPhase(html5parser.InitialPhase):
+    """ Consume XML Prologs """
+    def processComment(self, data):
+        if not data.startswith('?xml') or not data.endswith('?'):
+            html5parser.InitialPhase.processComment(self, data)
+
 class XmlRootPhase(html5parser.Phase):
+    """ Consume XML Prologs """
+    def processComment(self, data):
+        print repr(data)
+        if not data.startswith('?xml') or not data.endswith('?'):
+            html5parser.InitialPhase.processComment(self, data)
+
     """ Prime the Xml parser """
     def __getattr__(self, name):
         self.tree.openElements.append(self.tree.document)
diff --git a/tests/test_lxp.py b/tests/test_lxp.py
@@ -214,6 +214,17 @@ def test_title(self):
     self.assertXhtmlEquals("""<html xmlns="http://www.w3.org/1999/xhtml">
 <head><title>1 &lt; 2 &amp; 3</title></head>
 <body>
+</body></html>""")
+
+  def test_prolog(self):
+    self.assertXhtmlEquals("""<?xml version="1.0" encoding="UTF-8" ?>
+<html xmlns="http://www.w3.org/1999/xhtml">
+<head><title>PROLOG</title></head>
+<body>
+</body></html>""",
+"""<html xmlns="http://www.w3.org/1999/xhtml">
+<head><title>PROLOG</title></head>
+<body>
 </body></html>""")
 
 def buildTestSuite():
