DefaultUserAgent: use XMLDocumentBuilder with v.nu's HtmlParser instead of HtmlDocumentBuilder.

This is due to a DOCTYPE processing bug in HtmlDocumentBuilder.

Author: carlosame

src/io/sf/carte/doc/agent/net/DefaultUserAgent.java

@@ -20,8 +20,6 @@
 import java.nio.charset.StandardCharsets;
 import java.util.EnumSet;
 
-import javax.xml.parsers.DocumentBuilder;
-
 import org.w3c.dom.DocumentType;
 import org.w3c.dom.Element;
 import org.w3c.dom.NodeList;
@@ -39,7 +37,8 @@
 import io.sf.carte.doc.style.css.nsac.Parser;
 import io.sf.carte.doc.xml.dtd.DefaultEntityResolver;
 import io.sf.carte.util.agent.AgentUtil;
-import nu.validator.htmlparser.dom.HtmlDocumentBuilder;
+import nu.validator.htmlparser.common.XmlViolationPolicy;
+import nu.validator.htmlparser.sax.HtmlParser;
 
 /**
  * Default User Agent.
@@ -129,15 +128,15 @@ public DOMDocument readURL(URL url) throws IOException, io.sf.carte.doc.Document
 			}
 			isHtml = mimeType.equals("text/html");
 		}
-		DocumentBuilder builder;
+		XMLDocumentBuilder builder = new XMLDocumentBuilder(domImpl);
 		if (isHtml) {
-			builder = new HtmlDocumentBuilder(domImpl);
-			((HtmlDocumentBuilder) builder).setIgnoringComments(false);
+			HtmlParser parser = new HtmlParser(XmlViolationPolicy.ALTER_INFOSET);
+			parser.setReportingDoctype(true);
+			parser.setCommentPolicy(XmlViolationPolicy.ALLOW);
+			builder.setXMLReader(parser);
 		} else {
-			XMLDocumentBuilder xmlbuilder = new XMLDocumentBuilder(domImpl);
-			xmlbuilder.setIgnoreElementContentWhitespace(true);
-			xmlbuilder.setEntityResolver(resolver);
-			builder = xmlbuilder;
+			builder.setIgnoreElementContentWhitespace(true);
+			builder.setEntityResolver(resolver);
 		}
 		try {
 			is = openInputStream(con);