/**
* Uses a Vector of TransformerHandlers to pipe XML input document through a series of 1 or more
* transformations. Called by {@link #pipeDocument}.
*
* @param vTHandler Vector of Transformation Handlers (1 per stylesheet).
* @param source absolute URI to XML input
* @param target absolute path to transformation output.
*/
public void usePipe(Vector vTHandler, String source, String target)
throws TransformerException, TransformerConfigurationException, FileNotFoundException,
IOException, SAXException, SAXNotRecognizedException {
XMLReader reader = XMLReaderFactory.createXMLReader();
TransformerHandler tHFirst = (TransformerHandler) vTHandler.firstElement();
reader.setContentHandler(tHFirst);
reader.setProperty("http://xml.org/sax/properties/lexical-handler", tHFirst);
for (int i = 1; i < vTHandler.size(); i++) {
TransformerHandler tHFrom = (TransformerHandler) vTHandler.elementAt(i - 1);
TransformerHandler tHTo = (TransformerHandler) vTHandler.elementAt(i);
tHFrom.setResult(new SAXResult(tHTo));
}
TransformerHandler tHLast = (TransformerHandler) vTHandler.lastElement();
Transformer trans = tHLast.getTransformer();
Properties outputProps = trans.getOutputProperties();
Serializer serializer = SerializerFactory.getSerializer(outputProps);
FileOutputStream out = new FileOutputStream(target);
try {
serializer.setOutputStream(out);
tHLast.setResult(new SAXResult(serializer.asContentHandler()));
reader.parse(source);
} finally {
// Always clean up the FileOutputStream,
// even if an exception was thrown in the try block
if (out != null) out.close();
}
}
@Override
public boolean execute(Property inputProperty, Node outputNode, Context context)
throws Exception {
Binary binaryValue = inputProperty.getBinary();
CheckArg.isNotNull(binaryValue, "binary");
if (!outputNode.isNew()) {
outputNode = outputNode.addNode(XmlLexicon.DOCUMENT);
}
XmlSequencerHandler sequencingHandler = new XmlSequencerHandler(outputNode, scoping);
// Create the reader ...
XMLReader reader = XMLReaderFactory.createXMLReader();
reader.setContentHandler(sequencingHandler);
reader.setErrorHandler(sequencingHandler);
// Ensure handler acting as entity resolver 2
reader.setProperty(DECL_HANDLER_FEATURE, sequencingHandler);
// Ensure handler acting as lexical handler
reader.setProperty(LEXICAL_HANDLER_FEATURE, sequencingHandler);
// Ensure handler acting as entity resolver 2
setFeature(reader, ENTITY_RESOLVER_2_FEATURE, true);
// Prevent loading of external DTDs
setFeature(reader, LOAD_EXTERNAL_DTDS_FEATURE, false);
// Prevent the resolving of DTD entities into fully-qualified URIS
setFeature(reader, RESOLVE_DTD_URIS_FEATURE, false);
// Parse XML document
try (InputStream stream = binaryValue.getStream()) {
reader.parse(new InputSource(stream));
}
return true;
}
private void updateWiretap() {
if (contentHandler != null) {
if (wiretapContentHander != null) {
wrappedReader.setContentHandler(
new CombineContentHandler(wiretapContentHander, contentHandler));
} else {
wrappedReader.setContentHandler(contentHandler);
}
} else {
wrappedReader.setContentHandler(wiretapContentHander);
}
try {
if (lexicalHandler != null) {
if (wiretapLexicalHandler != null) {
wrappedReader.setProperty(
"http://xml.org/sax/properties/lexical-handler",
new CombineLexicalHandler(wiretapLexicalHandler, lexicalHandler));
} else {
wrappedReader.setProperty(
"http://xml.org/sax/properties/lexical-handler", lexicalHandler);
}
} else {
wrappedReader.setProperty(
"http://xml.org/sax/properties/lexical-handler", wiretapLexicalHandler);
}
} catch (SAXNotRecognizedException e) {
} catch (SAXNotSupportedException e) {
}
}
/**
* Reads malformed XML from the InputStream original and returns a new InputStream which can be
* used to read a well-formed version of the input
*
* @param original original input
* @return an {@link InputStream} which can be used to read a well-formed version of the input XML
* @throws ParseException if an exception occurs while parsing the input
*/
public static InputStream xmlizeInputStream(InputStream original) throws ParseException {
try {
ByteArrayOutputStream out = new ByteArrayOutputStream();
HTMLSchema schema = new HTMLSchema();
XMLReader reader = new Parser();
// TODO walk through the javadoc and tune more settings
// see tagsoup javadoc for details
reader.setProperty(Parser.schemaProperty, schema);
reader.setFeature(Parser.bogonsEmptyFeature, false);
reader.setFeature(Parser.ignorableWhitespaceFeature, true);
reader.setFeature(Parser.ignoreBogonsFeature, false);
Writer writeger = new OutputStreamWriter(out);
XMLWriter x = new XMLWriter(writeger);
reader.setContentHandler(x);
InputSource s = new InputSource(original);
reader.parse(s);
return new ByteArrayInputStream(out.toByteArray());
} catch (SAXException e) {
throw new ParseException(R("PBadXML"), e);
} catch (IOException e) {
throw new ParseException(R("PBadXML"), e);
}
}
public Struct validate(InputSource xml) throws PageException {
CFMLEngine engine = CFMLEngineFactory.getInstance();
warnings = engine.getCreationUtil().createArray();
errors = engine.getCreationUtil().createArray();
fatals = engine.getCreationUtil().createArray();
try {
XMLReader parser = new XMLUtilImpl().createXMLReader("org.apache.xerces.parsers.SAXParser");
parser.setContentHandler(this);
parser.setErrorHandler(this);
parser.setEntityResolver(this);
parser.setFeature("http://xml.org/sax/features/validation", true);
parser.setFeature("http://apache.org/xml/features/validation/schema", true);
parser.setFeature("http://apache.org/xml/features/validation/schema-full-checking", true);
// if(!validateNamespace)
if (!Util.isEmpty(strSchema))
parser.setProperty(
"http://apache.org/xml/properties/schema/external-noNamespaceSchemaLocation",
strSchema);
parser.parse(xml);
} catch (SAXException e) {
} catch (IOException e) {
throw engine.getExceptionUtil().createXMLException(e.getMessage());
}
// result
Struct result = engine.getCreationUtil().createStruct();
result.setEL("warnings", warnings);
result.setEL("errors", errors);
result.setEL("fatalerrors", fatals);
result.setEL("status", engine.getCastUtil().toBoolean(!hasErrors));
release();
return result;
}
private XMLReader getXMLReader(ContentHandler contentHandler, ErrorHandler errorHandler)
throws ParserConfigurationException, SAXException {
// setup sax factory ; be sure just one instance!
SAXParserFactory saxFactory = SAXParserFactory.newInstance();
// Enable validation stuff
saxFactory.setValidating(true);
saxFactory.setNamespaceAware(true);
// Create xml reader
SAXParser saxParser = saxFactory.newSAXParser();
XMLReader xmlReader = saxParser.getXMLReader();
// Setup xmlreader
xmlReader.setProperty(
XMLReaderObjectFactory.APACHE_PROPERTIES_INTERNAL_GRAMMARPOOL, grammarPool);
xmlReader.setFeature(Namespaces.SAX_VALIDATION, true);
xmlReader.setFeature(Namespaces.SAX_VALIDATION_DYNAMIC, false);
xmlReader.setFeature(XMLReaderObjectFactory.APACHE_FEATURES_VALIDATION_SCHEMA, true);
xmlReader.setFeature(XMLReaderObjectFactory.APACHE_PROPERTIES_LOAD_EXT_DTD, true);
xmlReader.setFeature(Namespaces.SAX_NAMESPACES_PREFIXES, true);
xmlReader.setContentHandler(contentHandler);
xmlReader.setErrorHandler(errorHandler);
return xmlReader;
}
private void processCommentLines(File file) throws SAXException, IOException {
SAXParser parser = newSaxParser(false);
XMLReader xmlReader = parser.getXMLReader();
commentHandler = new CommentHandler();
xmlReader.setProperty("http://xml.org/sax/properties/lexical-handler", commentHandler);
parser.parse(FileUtils.openInputStream(file), commentHandler);
}
/** Default constructor of MapLinksReader class. */
public MapLinksReader() {
super();
map = new HashMap<String, Map<String, String>>();
ancestorList = new ArrayList<String>(INT_16);
matchList = new ArrayList<String>(INT_16);
indexEntries = new StringBuffer(INT_1024);
firstMatchElement = null;
lastMatchElement = new HashSet<String>();
level = 0;
match = false;
validHref = true;
needResolveEntity = false;
topicPath = null;
inputFile = null;
try {
reader = StringUtils.getXMLReader();
reader.setContentHandler(this);
reader.setProperty(LEXICAL_HANDLER_PROPERTY, this);
// Added by william on 2009-11-8 for ampbug:2893664 start
reader.setFeature("http://apache.org/xml/features/scanner/notify-char-refs", true);
reader.setFeature("http://apache.org/xml/features/scanner/notify-builtin-refs", true);
// Added by william on 2009-11-8 for ampbug:2893664 end
reader.setFeature("http://xml.org/sax/features/namespaces", false);
} catch (final Exception e) {
logger.logException(e);
}
}
/**
* This will create a SAX XMLReader capable of parsing a DTD and configure it so that the DTD
* parsing events are routed to the handlers registered onto this SAXOutputter.
*
* @return <code>XMLReader</code> a SAX2 parser.
* @throws JDOMException if no parser can be created.
*/
private XMLReader createDTDParser() throws JDOMException {
XMLReader parser = null;
// Get a parser instance
try {
parser = createParser();
} catch (Exception ex1) {
throw new JDOMException("Error in SAX parser allocation", ex1);
}
// Register handlers
if (this.getDTDHandler() != null) {
parser.setDTDHandler(this.getDTDHandler());
}
if (this.getEntityResolver() != null) {
parser.setEntityResolver(this.getEntityResolver());
}
if (this.getLexicalHandler() != null) {
try {
parser.setProperty(SAX_PROPERTY_LEXICAL_HANDLER, this.getLexicalHandler());
} catch (SAXException ex1) {
try {
parser.setProperty(SAX_PROPERTY_LEXICAL_HANDLER_ALT, this.getLexicalHandler());
} catch (SAXException ex2) {
// Forget it!
}
}
}
if (this.getDeclHandler() != null) {
try {
parser.setProperty(SAX_PROPERTY_DECLARATION_HANDLER, this.getDeclHandler());
} catch (SAXException ex1) {
try {
parser.setProperty(SAX_PROPERTY_DECLARATION_HANDLER_ALT, this.getDeclHandler());
} catch (SAXException ex2) {
// Forget it!
}
}
}
// Absorb errors as much as possible, per Laurent
parser.setErrorHandler(new DefaultHandler());
return parser;
}
/**
* @param name
* @param value
* @throws SAXNotRecognizedException
* @throws SAXNotSupportedException
* @see org.xml.sax.XMLReader#setProperty(java.lang.String, java.lang.Object)
*/
public void setProperty(String name, Object value)
throws SAXNotRecognizedException, SAXNotSupportedException {
if ("http://xml.org/sax/properties/lexical-handler".equals(name)) {
lexicalHandler = (LexicalHandler) value;
updateWiretap();
} else {
wrappedReader.setProperty(name, value);
}
}
public SimpleDocTypeParser() throws SAXException {
xmlReader = XMLReaderFactory.createXMLReader();
xmlReader.setContentHandler(this);
// LexicalHandler
xmlReader.setProperty("http://xml.org/sax/properties/lexical-handler", this);
// DeclHandler
xmlReader.setProperty("http://xml.org/sax/properties/declaration-handler", this);
// DTD
xmlReader.setFeature("http://xml.org/sax/features/resolve-dtd-uris", false);
// *skip* resolving entities like DTDs
xmlReader.setEntityResolver(new NoEntityResolver());
// xmlReader.setProperty(
// "http://xml.org/sax/properties/declaration-handler", dh);
}
public void parse(InputStream xml, OutputStream finf, String workingDirectory) throws Exception {
SAXParser saxParser = getParser();
SAXDocumentSerializer documentSerializer = getSerializer(finf);
XMLReader reader = saxParser.getXMLReader();
reader.setProperty("http://xml.org/sax/properties/lexical-handler", documentSerializer);
reader.setContentHandler(documentSerializer);
if (workingDirectory != null) {
reader.setEntityResolver(createRelativePathResolver(workingDirectory));
}
reader.parse(new InputSource(xml));
}
// Register handler directly with the incremental parser
public void setLexicalHandler(org.xml.sax.ext.LexicalHandler handler) {
// Not supported by all SAX2 parsers but should work in Xerces:
try {
// Typecast required in Xerces2; SAXParser doesn't inheret XMLReader
// %OPT% Cast at asignment?
((XMLReader) fIncrementalParser)
.setProperty("http://xml.org/sax/properties/lexical-handler", handler);
} catch (org.xml.sax.SAXNotRecognizedException e) {
// Nothing we can do about it
} catch (org.xml.sax.SAXNotSupportedException e) {
// Nothing we can do about it
}
}
/**
* Parse the input source into a set of modifications.
*
* @param is
* @return an array of type Modification
* @throws ParserConfigurationException
* @throws IOException
* @throws SAXException
*/
public Modification[] parse(InputSource is)
throws ParserConfigurationException, IOException, SAXException {
final XMLReader reader = broker.getBrokerPool().getParserPool().borrowXMLReader();
try {
reader.setProperty(Namespaces.SAX_LEXICAL_HANDLER, this);
reader.setFeature(Namespaces.SAX_NAMESPACES, true);
reader.setFeature(Namespaces.SAX_NAMESPACES_PREFIXES, false);
reader.setContentHandler(this);
reader.parse(is);
final Modification mods[] = new Modification[modifications.size()];
return modifications.toArray(mods);
} finally {
broker.getBrokerPool().getParserPool().returnXMLReader(reader);
}
}
/** Parse the file and write its transformed content to the Writer. */
public void writeContent() throws SAXException {
SaxHandler handler = new SaxHandler();
XMLReader reader = XMLReaderFactory.createXMLReader();
reader.setContentHandler(handler);
reader.setProperty("http://xml.org/sax/properties/lexical-handler", handler);
try {
InputStream is = new BufferedInputStream(new FileInputStream(xdoc));
reader.parse(new InputSource(is));
} catch (IOException ex) {
throw new RuntimeException(ex);
}
}
@Override
protected void doGet(final HttpServletRequest request, final HttpServletResponse response)
throws ServletException, IOException {
LOG.trace("{@method} request = {}", request);
final String requestUri = request.getRequestURI();
if (!requestUri.matches("^/[^/]+/.+")) {
response.sendError(500);
return;
}
try {
final StringBuilder pageUrlBuilder = new StringBuilder();
pageUrlBuilder.append("http:/").append(requestUri);
if (request.getQueryString() != null) {
pageUrlBuilder.append("?").append(request.getQueryString());
}
final URL pageUrl = new URL(pageUrlBuilder.toString());
LOG.debug("GET: {}", pageUrl);
final XMLReader reader = new Parser();
reader.setProperty(Parser.schemaProperty, new HTMLSchema());
final ByteArrayOutputStream bytesOut = new ByteArrayOutputStream();
Writer writer = null;
try {
writer = new OutputStreamWriter(bytesOut);
reader.setContentHandler(new XMLWriter(writer));
final InputSource source = new InputSource();
source.setByteStream(pageUrl.openStream());
reader.parse(source);
} finally {
Closeables.closeQuietly(writer);
}
InputStream bytesIn = null;
try {
bytesIn = new ByteArrayInputStream(bytesOut.toByteArray());
ByteStreams.copy(bytesIn, response.getOutputStream());
} finally {
Closeables.closeQuietly(bytesIn);
Closeables.closeQuietly(response.getOutputStream());
}
} catch (final Exception e) {
throw new IllegalStateException(request.getRequestURI(), e);
}
}
public Xv4htContentHandler(XMLReader xmlReader, String catalog) {
super();
this.catalog = catalog;
lexicalHandler = new Xv4htLexicalHandler();
try {
xmlReader.setProperty("http://xml.org/sax/properties/lexical-handler", lexicalHandler);
} catch (SAXNotRecognizedException e) {
System.err.println("err 1");
return;
} catch (SAXNotSupportedException e) {
System.err.println("err 2");
return;
}
errHandler = new Xv4htErrorHandler();
xmlReader.setErrorHandler(errHandler);
xmlReader.setEntityResolver(new Xv4htEntityResolver());
}
public void parse(java.io.Reader src) throws IOException {
try {
org.xml.sax.XMLReader reader = org.openide.xml.XMLUtil.createXMLReader(false, false);
reader.setContentHandler(this);
reader.setEntityResolver(this);
org.xml.sax.InputSource is = new org.xml.sax.InputSource(src);
try {
reader.setProperty("http://xml.org/sax/properties/lexical-handler", this); // NOI18N
} catch (SAXException sex) {
XMLSettingsSupport.err.warning(
"Warning: XML parser does not support lexical-handler feature."); // NOI18N
}
reader.parse(is);
} catch (SAXException ex) {
IOException ioe = new IOException();
ioe.initCause(ex);
throw ioe;
}
}
protected void installLexicalHandler() {
XMLReader parent = getParent();
if (parent == null) {
throw new NullPointerException("No parent for filter");
}
// try to register for lexical events
for (String lexicalHandlerName : LEXICAL_HANDLER_NAMES) {
try {
parent.setProperty(lexicalHandlerName, this);
break;
} catch (SAXNotRecognizedException ex) {
// ignore
} catch (SAXNotSupportedException ex) {
// ignore
}
}
}
/**
* Creates a new SAX parser for use within this instance.
*
* @return The newly created parser.
* @throws ParserConfigurationException If a parser of the given configuration cannot be created.
* @throws SAXException If something in general goes wrong when creating the parser.
* @throws SAXNotRecognizedException If the <code>XMLReader</code> does not recognize the lexical
* handler configuration option.
* @throws SAXNotSupportedException If the <code>XMLReader</code> does not support the lexical
* handler configuration option.
*/
private final SAXParser createParser(SAXParserFactory parserFactory)
throws ParserConfigurationException, SAXException, SAXNotRecognizedException,
SAXNotSupportedException {
// Initialize the parser.
final SAXParser parser = parserFactory.newSAXParser();
final XMLReader reader = parser.getXMLReader();
reader.setProperty("http://xml.org/sax/properties/lexical-handler", this); // $NON-NLS-1$
// disable DTD validation (bug 63625)
try {
// be sure validation is "off" or the feature to ignore DTD's will not apply
reader.setFeature("http://xml.org/sax/features/validation", false); // $NON-NLS-1$
reader.setFeature(
"http://apache.org/xml/features/nonvalidating/load-external-dtd", false); // $NON-NLS-1$
} catch (SAXNotRecognizedException e) {
// not a big deal if the parser does not recognize the features
} catch (SAXNotSupportedException e) {
// not a big deal if the parser does not support the features
}
return parser;
}
@Override
public String parseFulltext(String html) {
if (!judge(html)) return null;
Parser theParser = null;
HTMLSchema theSchema = null;
XMLReader r;
if (theParser == null) theParser = new Parser();
r = theParser;
theSchema = new HTMLSchema();
try {
r.setProperty(Parser.schemaProperty, theSchema);
r.setFeature(Parser.namespacesFeature, false);
} catch (SAXNotRecognizedException e) {
logger.fatal(e.getMessage());
return null;
} catch (SAXNotSupportedException e) {
logger.fatal(e.getMessage());
return null;
}
RejuvenationResearchParserHandler h = new RejuvenationResearchParserHandler();
r.setContentHandler(h);
InputStream fin = new ByteArrayInputStream(html.getBytes());
InputSource s = new InputSource(fin);
s.setEncoding("utf8");
try {
r.parse(s);
fin.close();
return h.getContent();
} catch (IOException e) {
logger.fatal(e.getMessage());
} catch (SAXException e) {
logger.fatal(e.getMessage());
}
return null;
}
private static void inputSourceToSAX(
InputSource inputSource,
XMLReceiver xmlReceiver,
XMLUtils.ParserConfiguration parserConfiguration,
boolean handleLexical) {
// Insert XInclude processor if needed
final TransformerURIResolver resolver;
if (parserConfiguration.handleXInclude) {
parserConfiguration =
new XMLUtils.ParserConfiguration(
parserConfiguration.validating,
false,
parserConfiguration.externalEntities,
parserConfiguration.uriReferences);
resolver = new TransformerURIResolver(XMLUtils.ParserConfiguration.PLAIN);
xmlReceiver =
new XIncludeReceiver(null, xmlReceiver, parserConfiguration.uriReferences, resolver);
} else {
resolver = null;
}
try {
final XMLReader xmlReader = newSAXParser(parserConfiguration).getXMLReader();
xmlReader.setContentHandler(xmlReceiver);
if (handleLexical) xmlReader.setProperty(XMLConstants.SAX_LEXICAL_HANDLER, xmlReceiver);
xmlReader.setEntityResolver(ENTITY_RESOLVER);
xmlReader.setErrorHandler(ERROR_HANDLER);
xmlReader.parse(inputSource);
} catch (SAXParseException e) {
throw new ValidationException(e.getMessage(), new LocationData(e));
} catch (Exception e) {
throw new OXFException(e);
} finally {
if (resolver != null) resolver.destroy();
}
}
/**
* Retrieves a cached XMLReader for this thread, or creates a new XMLReader, if the existing
* reader is in use. When the caller no longer needs the reader, it must release it with a call to
* {@link #releaseXMLReader}.
*/
public synchronized XMLReader getXMLReader() throws SAXException {
XMLReader reader;
if (m_readers == null) {
// When the m_readers.get() method is called for the first time
// on a thread, a new XMLReader will automatically be created.
m_readers = new ThreadLocal();
}
if (m_inUse == null) {
m_inUse = new HashMap();
}
// If the cached reader for this thread is in use, construct a new
// one; otherwise, return the cached reader unless it isn't an
// instance of the class set in the 'org.xml.sax.driver' property
reader = (XMLReader) m_readers.get();
boolean threadHasReader = (reader != null);
String factory = SecuritySupport.getSystemProperty(property);
if (threadHasReader
&& m_inUse.get(reader) != Boolean.TRUE
&& (factory == null || reader.getClass().getName().equals(factory))) {
m_inUse.put(reader, Boolean.TRUE);
} else {
try {
try {
// According to JAXP 1.2 specification, if a SAXSource
// is created using a SAX InputSource the Transformer or
// TransformerFactory creates a reader via the
// XMLReaderFactory if setXMLReader is not used
reader = XMLReaderFactory.createXMLReader();
try {
reader.setFeature(XMLConstants.FEATURE_SECURE_PROCESSING, _secureProcessing);
} catch (SAXNotRecognizedException e) {
System.err.println("Warning: " + reader.getClass().getName() + ": " + e.getMessage());
}
} catch (Exception e) {
try {
// If unable to create an instance, let's try to use
// the XMLReader from JAXP
if (m_parserFactory == null) {
m_parserFactory = FactoryImpl.getSAXFactory(m_useServicesMechanism);
m_parserFactory.setNamespaceAware(true);
}
reader = m_parserFactory.newSAXParser().getXMLReader();
} catch (ParserConfigurationException pce) {
throw pce; // pass along pce
}
}
try {
reader.setFeature(NAMESPACES_FEATURE, true);
reader.setFeature(NAMESPACE_PREFIXES_FEATURE, false);
} catch (SAXException se) {
// Try to carry on if we've got a parser that
// doesn't know about namespace prefixes.
}
} catch (ParserConfigurationException ex) {
throw new SAXException(ex);
} catch (FactoryConfigurationError ex1) {
throw new SAXException(ex1.toString());
} catch (NoSuchMethodError ex2) {
} catch (AbstractMethodError ame) {
}
// Cache the XMLReader if this is the first time we've created
// a reader for this thread.
if (!threadHasReader) {
m_readers.set(reader);
m_inUse.put(reader, Boolean.TRUE);
}
}
try {
// reader is cached, but this property might have been reset
reader.setProperty(XMLConstants.ACCESS_EXTERNAL_DTD, _accessExternalDTD);
} catch (SAXException se) {
System.err.println("Warning: " + reader.getClass().getName() + ": " + se.getMessage());
}
try {
if (_xmlSecurityManager != null) {
for (XMLSecurityManager.Limit limit : XMLSecurityManager.Limit.values()) {
reader.setProperty(limit.apiProperty(), _xmlSecurityManager.getLimitValueAsString(limit));
}
if (_xmlSecurityManager.printEntityCountInfo()) {
reader.setProperty(XalanConstants.JDK_ENTITY_COUNT_INFO, XalanConstants.JDK_YES);
}
}
} catch (SAXException se) {
System.err.println("Warning: " + reader.getClass().getName() + ": " + se.getMessage());
}
return reader;
}
/**
* Converts an HTML document to XML.
*
* @param io io reference
* @param opts html options
* @return parser
* @throws IOException I/O exception
*/
private static IO toXML(final IO io, final HtmlOptions opts) throws IOException {
// reader could not be initialized; fall back to XML
if (READER == null) return io;
try {
// tries to extract the encoding from the input
final TextInput ti = new TextInput(io);
String enc = ti.encoding();
final byte[] content = ti.content();
// looks for a charset definition
final byte[] encoding = token("charset=");
int cs = indexOf(content, encoding);
if (cs > 0) {
// extracts the encoding string
cs += encoding.length;
int ce = cs;
final int cl = content.length;
while (++ce < cl && content[ce] > 0x28) ;
enc = string(substring(content, cs, ce));
}
// define input
final InputSource is = new InputSource(new ArrayInput(content));
is.setEncoding(supported(enc) ? normEncoding(enc) : UTF8);
// define output
final StringWriter sw = new StringWriter();
final XMLReader reader = (XMLReader) Reflect.get(READER);
final Object writer = Reflect.get(WRITER, sw);
// set TagSoup options
if (opts.get(HtmlOptions.HTML)) {
reader.setFeature("http://xml.org/sax/features/namespaces", false);
opt("method", "html");
opt("omit-xml-declaration", "yes");
}
if (opts.get(HtmlOptions.NONS))
reader.setFeature("http://xml.org/sax/features/namespaces", false);
if (opts.get(HtmlOptions.OMITXML)) opt("omit-xml-declaration", "yes");
if (opts.get(HtmlOptions.NOBOGONS)) reader.setFeature(FEATURES + "ignore-bogons", true);
if (opts.get(HtmlOptions.NODEFAULTS))
reader.setFeature(FEATURES + "default-attributes", false);
if (opts.get(HtmlOptions.NOCOLONS)) reader.setFeature(FEATURES + "translate-colons", true);
if (opts.get(HtmlOptions.NORESTART)) reader.setFeature(FEATURES + "restart-elements", false);
if (opts.get(HtmlOptions.IGNORABLE))
reader.setFeature(FEATURES + "ignorable-whitespace", true);
if (opts.get(HtmlOptions.EMPTYBOGONS)) reader.setFeature(FEATURES + "bogons-empty", true);
if (opts.get(HtmlOptions.ANY)) reader.setFeature(FEATURES + "bogons-empty", false);
if (opts.get(HtmlOptions.NOROOTBOGONS)) reader.setFeature(FEATURES + "root-bogons", false);
if (opts.get(HtmlOptions.NOCDATA)) reader.setFeature(FEATURES + "cdata-elements", false);
if (opts.get(HtmlOptions.LEXICAL))
reader.setProperty("http://xml.org/sax/properties/lexical-handler", writer);
if (opts.contains(HtmlOptions.METHOD)) opt("method", opts.get(HtmlOptions.METHOD));
if (opts.contains(HtmlOptions.DOCTYPESYS))
opt("doctype-system", opts.get(HtmlOptions.DOCTYPESYS));
if (opts.contains(HtmlOptions.DOCTYPEPUB))
opt("doctype-public", opts.get(HtmlOptions.DOCTYPEPUB));
if (opts.contains(HtmlOptions.ENCODING)) is.setEncoding(opts.get(HtmlOptions.ENCODING));
// end TagSoup options
reader.setContentHandler((ContentHandler) writer);
reader.parse(is);
return new IOContent(token(sw.toString()), io.name());
} catch (final SAXException ex) {
Util.errln(ex);
return io;
}
}
/**
* Validate XML data from reader using specified grammar.
*
* @param grammarUrl User supplied path to grammar.
* @param stream XML input.
* @return Validation report containing all validation info.
*/
public ValidationReport validateParse(InputStream stream, String grammarUrl) {
logger.debug("Start validation.");
ValidationReport report = new ValidationReport();
ValidationContentHandler contenthandler = new ValidationContentHandler();
try {
XMLReader xmlReader = getXMLReader(contenthandler, report);
if (grammarUrl == null) {
// Scenario 1 : no params - use system catalog
logger.debug("Validation using system catalog.");
xmlReader.setProperty(
XMLReaderObjectFactory.APACHE_PROPERTIES_ENTITYRESOLVER, systemCatalogResolver);
} else if (grammarUrl.endsWith(".xml")) {
// Scenario 2 : path to catalog (xml)
logger.debug("Validation using user specified catalog '" + grammarUrl + "'.");
eXistXMLCatalogResolver resolver = new eXistXMLCatalogResolver();
resolver.setCatalogList(new String[] {grammarUrl});
xmlReader.setProperty(XMLReaderObjectFactory.APACHE_PROPERTIES_ENTITYRESOLVER, resolver);
} else if (grammarUrl.endsWith("/")) {
// Scenario 3 : path to collection ("/"): search.
logger.debug("Validation using searched grammar, start from '" + grammarUrl + "'.");
SearchResourceResolver resolver = new SearchResourceResolver(grammarUrl, brokerPool);
xmlReader.setProperty(XMLReaderObjectFactory.APACHE_PROPERTIES_ENTITYRESOLVER, resolver);
} else {
// Scenario 4 : path to grammar (xsd, dtd) specified.
logger.debug("Validation using specified grammar '" + grammarUrl + "'.");
AnyUriResolver resolver = new AnyUriResolver(grammarUrl);
xmlReader.setProperty(XMLReaderObjectFactory.APACHE_PROPERTIES_ENTITYRESOLVER, resolver);
}
logger.debug("Validation started.");
report.start();
InputSource source = new InputSource(stream);
xmlReader.parse(source);
logger.debug("Validation stopped.");
report.stop();
report.setNamespaceUri(contenthandler.getNamespaceUri());
if (!report.isValid()) {
logger.debug("Document is not valid.");
}
} catch (ExistIOException ex) {
logger.error(ex.getCause());
report.setThrowable(ex.getCause());
} catch (Exception ex) {
logger.error(ex);
report.setThrowable(ex);
} finally {
report.stop();
logger.debug("Validation performed in " + report.getValidationDuration() + " msec.");
}
return report;
}
public static void setLexicalHandler(XMLReader xmlReader, LexicalHandler handler)
throws SAXNotRecognizedException, SAXNotSupportedException {
logger.debug("setLexicalHandler(xmlReader={}, handler={}) - start", xmlReader, handler);
xmlReader.setProperty(LEXICAL_HANDLER_PROPERTY_NAME, handler);
}
public static void main(String args[]) {
// Printing arguments just to show them
boolean debug = false;
if (debug) {
for (int i = 0; i < args.length; i++) {
System.err.printf("arg[%d]: %s\n", i, args[i]);
}
}
// Now the real program
InputStream in = System.in;
PrintStream out = System.out;
PrintStream err = System.err;
String usage =
"Usage: "
+ OLEWrapper.class.getName()
+ " [file]\n"
+ "\tfile = input file to read, default is stdin";
if (args.length > 1) {
out.println("Error: more than 1 argument");
out.println(usage);
System.exit(1);
} else if (args.length == 1) {
try {
in = new BufferedInputStream(new FileInputStream(args[0]));
} catch (FileNotFoundException e) {
err.println("file " + args[0] + "was not found.");
System.exit(2);
}
}
// Then do the work...
SAXParserFactory factory = SAXParserFactory.newInstance();
factory.setNamespaceAware(true);
try {
SAXParser saxParser = factory.newSAXParser();
DefaultHandler handler = new OLEImportHandler(out, err);
XMLReader reader = saxParser.getXMLReader();
if (debug) {
out.println("factory.isNamespaceAware()=" + factory.isNamespaceAware());
out.println("saxParser.isNamespaceAware()=" + saxParser.isNamespaceAware());
out.println(SaxNSURI + "=" + reader.getFeature(SaxNSURI));
System.exit(0);
}
reader.setContentHandler(handler);
// Register lexical handler for comment processing
reader.setProperty(SaxLexHandlerURI, handler);
reader.parse(new InputSource(in));
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (ParserConfigurationException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (SAXException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
System.exit(0);
}
@Override
public void setProperty(String name, Object value)
throws SAXNotRecognizedException, SAXNotSupportedException {
xmlReader.setProperty(name, value);
}
public void validate(Source source, Result result) throws SAXException, IOException {
if (result instanceof SAXResult || result == null) {
final SAXSource saxSource = (SAXSource) source;
final SAXResult saxResult = (SAXResult) result;
if (result != null) {
setContentHandler(saxResult.getHandler());
}
try {
XMLReader reader = saxSource.getXMLReader();
if (reader == null) {
// create one now
SAXParserFactory spf = SAXParserFactory.newInstance();
spf.setNamespaceAware(true);
try {
reader = spf.newSAXParser().getXMLReader();
// If this is a Xerces SAX parser, set the security manager if there is one
if (reader instanceof com.sun.org.apache.xerces.internal.parsers.SAXParser) {
SecurityManager securityManager =
(SecurityManager) fComponentManager.getProperty(SECURITY_MANAGER);
if (securityManager != null) {
try {
reader.setProperty(SECURITY_MANAGER, securityManager);
}
// Ignore the exception if the security manager cannot be set.
catch (SAXException exc) {
}
}
}
} catch (Exception e) {
// this is impossible, but better safe than sorry
throw new FactoryConfigurationError(e);
}
}
// If XML names and Namespace URIs are already internalized we
// can avoid running them through the SymbolTable.
try {
fStringsInternalized = reader.getFeature(STRING_INTERNING);
} catch (SAXException exc) {
// The feature isn't recognized or getting it is not supported.
// In either case, assume that strings are not internalized.
fStringsInternalized = false;
}
ErrorHandler errorHandler = fComponentManager.getErrorHandler();
reader.setErrorHandler(
errorHandler != null ? errorHandler : DraconianErrorHandler.getInstance());
reader.setEntityResolver(fResolutionForwarder);
fResolutionForwarder.setEntityResolver(fComponentManager.getResourceResolver());
reader.setContentHandler(this);
reader.setDTDHandler(this);
InputSource is = saxSource.getInputSource();
reader.parse(is);
} finally {
// release the reference to user's handler ASAP
setContentHandler(null);
}
return;
}
throw new IllegalArgumentException(
JAXPValidationMessageFormatter.formatMessage(
Locale.getDefault(),
"SourceResultMismatch",
new Object[] {source.getClass().getName(), result.getClass().getName()}));
}
/**
* Get an instance of a DTM, loaded with the content from the specified source. If the unique flag
* is true, a new instance will always be returned. Otherwise it is up to the DTMManager to return
* a new instance or an instance that it already created and may be being used by someone else. (I
* think more parameters will need to be added for error handling, and entity resolution).
*
* @param source the specification of the source object.
* @param unique true if the returned DTM must be unique, probably because it is going to be
* mutated.
* @param whiteSpaceFilter Enables filtering of whitespace nodes, and may be null.
* @param incremental true if the DTM should be built incrementally, if possible.
* @param doIndexing true if the caller considers it worth it to use indexing schemes.
* @param hasUserReader true if <code>source</code> is a <code>SAXSource</code> object that has an
* <code>XMLReader</code>, that was specified by the user.
* @param size Specifies initial size of tables that represent the DTM
* @param buildIdIndex true if the id index table should be built.
* @param newNameTable true if we want to use a separate ExpandedNameTable for this DTM.
* @return a non-null DTM reference.
*/
public DTM getDTM(
Source source,
boolean unique,
DTMWSFilter whiteSpaceFilter,
boolean incremental,
boolean doIndexing,
boolean hasUserReader,
int size,
boolean buildIdIndex,
boolean newNameTable) {
if (DEBUG && null != source) {
System.out.println(
"Starting " + (unique ? "UNIQUE" : "shared") + " source: " + source.getSystemId());
}
int dtmPos = getFirstFreeDTMID();
int documentID = dtmPos << IDENT_DTM_NODE_BITS;
if ((null != source) && source instanceof StAXSource) {
final StAXSource staxSource = (StAXSource) source;
StAXEvent2SAX staxevent2sax = null;
StAXStream2SAX staxStream2SAX = null;
if (staxSource.getXMLEventReader() != null) {
final XMLEventReader xmlEventReader = staxSource.getXMLEventReader();
staxevent2sax = new StAXEvent2SAX(xmlEventReader);
} else if (staxSource.getXMLStreamReader() != null) {
final XMLStreamReader xmlStreamReader = staxSource.getXMLStreamReader();
staxStream2SAX = new StAXStream2SAX(xmlStreamReader);
}
SAXImpl dtm;
if (size <= 0) {
dtm =
new SAXImpl(
this,
source,
documentID,
whiteSpaceFilter,
null,
doIndexing,
DTMDefaultBase.DEFAULT_BLOCKSIZE,
buildIdIndex,
newNameTable);
} else {
dtm =
new SAXImpl(
this,
source,
documentID,
whiteSpaceFilter,
null,
doIndexing,
size,
buildIdIndex,
newNameTable);
}
dtm.setDocumentURI(source.getSystemId());
addDTM(dtm, dtmPos, 0);
try {
if (staxevent2sax != null) {
staxevent2sax.setContentHandler(dtm);
staxevent2sax.parse();
} else if (staxStream2SAX != null) {
staxStream2SAX.setContentHandler(dtm);
staxStream2SAX.parse();
}
} catch (RuntimeException re) {
throw re;
} catch (Exception e) {
throw new com.sun.org.apache.xml.internal.utils.WrappedRuntimeException(e);
}
return dtm;
} else if ((null != source) && source instanceof DOMSource) {
final DOMSource domsrc = (DOMSource) source;
final org.w3c.dom.Node node = domsrc.getNode();
final DOM2SAX dom2sax = new DOM2SAX(node);
SAXImpl dtm;
if (size <= 0) {
dtm =
new SAXImpl(
this,
source,
documentID,
whiteSpaceFilter,
null,
doIndexing,
DTMDefaultBase.DEFAULT_BLOCKSIZE,
buildIdIndex,
newNameTable);
} else {
dtm =
new SAXImpl(
this,
source,
documentID,
whiteSpaceFilter,
null,
doIndexing,
size,
buildIdIndex,
newNameTable);
}
dtm.setDocumentURI(source.getSystemId());
addDTM(dtm, dtmPos, 0);
dom2sax.setContentHandler(dtm);
try {
dom2sax.parse();
} catch (RuntimeException re) {
throw re;
} catch (Exception e) {
throw new com.sun.org.apache.xml.internal.utils.WrappedRuntimeException(e);
}
return dtm;
} else {
boolean isSAXSource = (null != source) ? (source instanceof SAXSource) : true;
boolean isStreamSource = (null != source) ? (source instanceof StreamSource) : false;
if (isSAXSource || isStreamSource) {
XMLReader reader;
InputSource xmlSource;
if (null == source) {
xmlSource = null;
reader = null;
hasUserReader = false; // Make sure the user didn't lie
} else {
reader = getXMLReader(source);
xmlSource = SAXSource.sourceToInputSource(source);
String urlOfSource = xmlSource.getSystemId();
if (null != urlOfSource) {
try {
urlOfSource = SystemIDResolver.getAbsoluteURI(urlOfSource);
} catch (Exception e) {
// %REVIEW% Is there a better way to send a warning?
System.err.println("Can not absolutize URL: " + urlOfSource);
}
xmlSource.setSystemId(urlOfSource);
}
}
// Create the basic SAX2DTM.
SAXImpl dtm;
if (size <= 0) {
dtm =
new SAXImpl(
this,
source,
documentID,
whiteSpaceFilter,
null,
doIndexing,
DTMDefaultBase.DEFAULT_BLOCKSIZE,
buildIdIndex,
newNameTable);
} else {
dtm =
new SAXImpl(
this,
source,
documentID,
whiteSpaceFilter,
null,
doIndexing,
size,
buildIdIndex,
newNameTable);
}
// Go ahead and add the DTM to the lookup table. This needs to be
// done before any parsing occurs. Note offset 0, since we've just
// created a new DTM.
addDTM(dtm, dtmPos, 0);
if (null == reader) {
// Then the user will construct it themselves.
return dtm;
}
reader.setContentHandler(dtm.getBuilder());
if (!hasUserReader || null == reader.getDTDHandler()) {
reader.setDTDHandler(dtm);
}
if (!hasUserReader || null == reader.getErrorHandler()) {
reader.setErrorHandler(dtm);
}
try {
reader.setProperty("http://xml.org/sax/properties/lexical-handler", dtm);
} catch (SAXNotRecognizedException e) {
} catch (SAXNotSupportedException e) {
}
try {
reader.parse(xmlSource);
} catch (RuntimeException re) {
throw re;
} catch (Exception e) {
throw new com.sun.org.apache.xml.internal.utils.WrappedRuntimeException(e);
} finally {
if (!hasUserReader) {
releaseXMLReader(reader);
}
}
if (DUMPTREE) {
System.out.println("Dumping SAX2DOM");
dtm.dumpDTM(System.err);
}
return dtm;
} else {
// It should have been handled by a derived class or the caller
// made a mistake.
throw new DTMException(
XMLMessages.createXMLMessage(
XMLErrorResources.ER_NOT_SUPPORTED, new Object[] {source}));
}
}
}
