All Downloads are FREE. Search and download functionalities are using the official Maven repository.

com.sun.tools.ws.wsdl.parser.DOMForest Maven / Gradle / Ivy

Go to download

Open source Reference Implementation of JSR-224: Java API for XML Web Services

There is a newer version: 4.0.3
Show newest version
/*
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS HEADER.
 *
 * Copyright (c) 1997-2010 Oracle and/or its affiliates. All rights reserved.
 *
 * The contents of this file are subject to the terms of either the GNU
 * General Public License Version 2 only ("GPL") or the Common Development
 * and Distribution License("CDDL") (collectively, the "License").  You
 * may not use this file except in compliance with the License.  You can
 * obtain a copy of the License at
 * https://glassfish.dev.java.net/public/CDDL+GPL_1_1.html
 * or packager/legal/LICENSE.txt.  See the License for the specific
 * language governing permissions and limitations under the License.
 *
 * When distributing the software, include this License Header Notice in each
 * file and include the License file at packager/legal/LICENSE.txt.
 *
 * GPL Classpath Exception:
 * Oracle designates this particular file as subject to the "Classpath"
 * exception as provided by Oracle in the GPL Version 2 section of the License
 * file that accompanied this code.
 *
 * Modifications:
 * If applicable, add the following below the License Header, with the fields
 * enclosed by brackets [] replaced by your own identifying information:
 * "Portions Copyright [year] [name of copyright owner]"
 *
 * Contributor(s):
 * If you wish your version of this file to be governed by only the CDDL or
 * only the GPL Version 2, indicate your decision by adding "[Contributor]
 * elects to include this software in this distribution under the [CDDL or GPL
 * Version 2] license."  If you don't indicate a single choice of license, a
 * recipient has the option to distribute your version of this file under
 * either the CDDL, the GPL Version 2 or to extend the choice of license to
 * its licensees as provided above.  However, if you add GPL Version 2 code
 * and therefore, elected the GPL Version 2 license, then the option applies
 * only if the new code is made subject to such option by the copyright
 * holder.
 */

package com.sun.tools.ws.wsdl.parser;

import com.sun.istack.NotNull;
import com.sun.tools.ws.resources.WscompileMessages;
import com.sun.tools.ws.wscompile.AbortException;
import com.sun.tools.ws.wscompile.DefaultAuthenticator;
import com.sun.tools.ws.wscompile.ErrorReceiver;
import com.sun.tools.ws.wscompile.WsimportOptions;
import com.sun.tools.ws.wsdl.document.schema.SchemaConstants;
import com.sun.tools.xjc.reader.internalizer.LocatorTable;
import com.sun.xml.bind.marshaller.DataWriter;
import com.sun.xml.ws.util.JAXWSUtils;
import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.NodeList;
import org.xml.sax.ContentHandler;
import org.xml.sax.*;
import org.xml.sax.helpers.XMLFilterImpl;

import javax.xml.parsers.DocumentBuilder;
import javax.xml.parsers.DocumentBuilderFactory;
import javax.xml.parsers.ParserConfigurationException;
import javax.xml.parsers.SAXParserFactory;
import javax.xml.transform.Transformer;
import javax.xml.transform.TransformerException;
import javax.xml.transform.TransformerFactory;
import javax.xml.transform.dom.DOMSource;
import javax.xml.transform.sax.SAXResult;
import javax.net.ssl.HttpsURLConnection;
import javax.net.ssl.HostnameVerifier;
import javax.net.ssl.SSLSession;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.io.OutputStreamWriter;
import java.net.*;
import java.util.*;

/**
 * @author Vivek Pandey
 */
public class DOMForest {
    /**
     * To correctly feed documents to a schema parser, we need to remember
     * which documents (of the forest) were given as the root
     * documents, and which of them are read as included/imported
     * documents.
     * 

*

* Set of system ids as strings. */ protected final Set rootDocuments = new HashSet(); /** * Contains wsdl:import(s) */ protected final Set externalReferences = new HashSet(); /** * actual data storage map<SystemId,Document>. */ protected final Map core = new HashMap(); protected final ErrorReceiver errorReceiver; private final DocumentBuilder documentBuilder; private final SAXParserFactory parserFactory; /** * inlined schema elements inside wsdl:type section */ protected final List inlinedSchemaElements = new ArrayList(); /** * Stores location information for all the trees in this forest. */ public final LocatorTable locatorTable = new LocatorTable(); protected final EntityResolver entityResolver; /** * Stores all the outer-most <jaxb:bindings> customizations. */ public final Set outerMostBindings = new HashSet(); /** * Schema language dependent part of the processing. */ protected final InternalizationLogic logic; protected final WsimportOptions options; public DOMForest(InternalizationLogic logic, @NotNull EntityResolver entityResolver, WsimportOptions options, ErrorReceiver errReceiver) { this.options = options; this.entityResolver = entityResolver; this.errorReceiver = errReceiver; this.logic = logic; try { DocumentBuilderFactory dbf = DocumentBuilderFactory.newInstance(); dbf.setNamespaceAware(true); this.documentBuilder = dbf.newDocumentBuilder(); this.parserFactory = SAXParserFactory.newInstance(); this.parserFactory.setNamespaceAware(true); } catch (ParserConfigurationException e) { throw new AssertionError(e); } } public List getInlinedSchemaElement() { return inlinedSchemaElements; } public @NotNull Document parse(InputSource source, boolean root) throws SAXException, IOException { if (source.getSystemId() == null) throw new IllegalArgumentException(); return parse(source.getSystemId(), source, root); } /** * Parses an XML at the given location ( * and XMLs referenced by it) into DOM trees * and stores them to this forest. * * @return the parsed DOM document object. */ public Document parse(String systemId, boolean root) throws SAXException, IOException{ systemId = normalizeSystemId(systemId); InputSource is = null; // allow entity resolver to find the actual byte stream. is = entityResolver.resolveEntity(null, systemId); if (is == null) is = new InputSource(systemId); else { resolvedCache.put(systemId, is.getSystemId()); systemId=is.getSystemId(); } if (core.containsKey(systemId)) { // this document has already been parsed. Just ignore. return core.get(systemId); } if(!root) addExternalReferences(systemId); // but we still use the original system Id as the key. return parse(systemId, is, root); } protected Map resolvedCache = new HashMap(); public Map getReferencedEntityMap() { return resolvedCache; } /** * Parses the given document and add it to the DOM forest. * * @return null if there was a parse error. otherwise non-null. */ private @NotNull Document parse(String systemId, InputSource inputSource, boolean root) throws SAXException, IOException{ Document dom = documentBuilder.newDocument(); systemId = normalizeSystemId(systemId); // put into the map before growing a tree, to // prevent recursive reference from causing infinite loop. core.put(systemId, dom); dom.setDocumentURI(systemId); if (root) rootDocuments.add(systemId); try { XMLReader reader = createReader(dom); InputStream is = null; if(inputSource.getByteStream() == null){ inputSource = entityResolver.resolveEntity(null, systemId); } reader.parse(inputSource); Element doc = dom.getDocumentElement(); if (doc == null) { return null; } NodeList schemas = doc.getElementsByTagNameNS(SchemaConstants.NS_XSD, "schema"); for (int i = 0; i < schemas.getLength(); i++) { inlinedSchemaElements.add((Element) schemas.item(i)); } } catch (ParserConfigurationException e) { errorReceiver.error(e); throw new SAXException(e.getMessage()); } resolvedCache.put(systemId, dom.getDocumentURI()); return dom; } public void addExternalReferences(String ref) { if (!externalReferences.contains(ref)) externalReferences.add(ref); } public Set getExternalReferences() { return externalReferences; } public interface Handler extends ContentHandler { /** * Gets the DOM that was built. */ public Document getDocument(); } /** * Returns a {@link org.xml.sax.XMLReader} to parse a document into this DOM forest. *

* This version requires that the DOM object to be created and registered * to the map beforehand. */ private XMLReader createReader(Document dom) throws SAXException, ParserConfigurationException { XMLReader reader = parserFactory.newSAXParser().getXMLReader(); DOMBuilder dombuilder = new DOMBuilder(dom, locatorTable, outerMostBindings); try { reader.setProperty("http://xml.org/sax/properties/lexical-handler", dombuilder); } catch(SAXException e) { errorReceiver.debug(e.getMessage()); } ContentHandler handler = new WhitespaceStripper(dombuilder, errorReceiver, entityResolver); handler = new VersionChecker(handler, errorReceiver, entityResolver); // insert the reference finder so that // included/imported schemas will be also parsed XMLFilterImpl f = logic.createExternalReferenceFinder(this); f.setContentHandler(handler); if (errorReceiver != null) f.setErrorHandler(errorReceiver); f.setEntityResolver(entityResolver); reader.setContentHandler(f); if (errorReceiver != null) reader.setErrorHandler(errorReceiver); reader.setEntityResolver(entityResolver); return reader; } private String normalizeSystemId(String systemId) { try { systemId = new URI(systemId).normalize().toString(); } catch (URISyntaxException e) { // leave the system ID untouched. In my experience URI is often too strict } return systemId; } boolean isExtensionMode() { return options.isExtensionMode(); } /** * Gets the DOM tree associated with the specified system ID, * or null if none is found. */ public Document get(String systemId) { Document doc = core.get(systemId); if (doc == null && systemId.startsWith("file:/") && !systemId.startsWith("file://")) { // As of JDK1.4, java.net.URL.toExternal method returns URLs like // "file:/abc/def/ghi" which is an incorrect file protocol URL according to RFC1738. // Some other correctly functioning parts return the correct URLs ("file:///abc/def/ghi"), // and this descripancy breaks DOM look up by system ID. // this extra check solves this problem. doc = core.get("file://" + systemId.substring(5)); } if (doc == null && systemId.startsWith("file:")) { // on Windows, filenames are case insensitive. // perform case-insensitive search for improved user experience String systemPath = getPath(systemId); for (String key : core.keySet()) { if (key.startsWith("file:") && getPath(key).equalsIgnoreCase(systemPath)) { doc = core.get(key); break; } } } return doc; } /** * Strips off the leading 'file:///' portion from an URL. */ private String getPath(String key) { key = key.substring(5); // skip 'file:' while (key.length() > 0 && key.charAt(0) == '/') key = key.substring(1); return key; } /** * Gets all the system IDs of the documents. */ public String[] listSystemIDs() { return core.keySet().toArray(new String[core.keySet().size()]); } /** * Gets the system ID from which the given DOM is parsed. *

* Poor-man's base URI. */ public String getSystemId(Document dom) { for (Map.Entry e : core.entrySet()) { if (e.getValue() == dom) return e.getKey(); } return null; } /** * Gets the first one (which is more or less random) in {@link #rootDocuments}. */ public String getFirstRootDocument() { if(rootDocuments.isEmpty()) return null; return rootDocuments.iterator().next(); } public Set getRootDocuments() { return rootDocuments; } /** * Dumps the contents of the forest to the specified stream. *

* This is a debug method. As such, error handling is sloppy. */ public void dump(OutputStream out) throws IOException { try { // create identity transformer Transformer it = TransformerFactory.newInstance().newTransformer(); for (Map.Entry e : core.entrySet()) { out.write(("---<< " + e.getKey() + '\n').getBytes()); DataWriter dw = new DataWriter(new OutputStreamWriter(out), null); dw.setIndentStep(" "); it.transform(new DOMSource(e.getValue()), new SAXResult(dw)); out.write("\n\n\n".getBytes()); } } catch (TransformerException e) { e.printStackTrace(); } } }





© 2015 - 2024 Weber Informatics LLC | Privacy Policy