Merge pull request #659 from dubinsky/retain-systemId

Retain SystemId

Author: dubinsky

jvm/src/test/scala/scala/xml/XMLTest.scala

@@ -24,10 +24,10 @@ class XMLTestJVM {
   def Elem(prefix: String, label: String, attributes: MetaData, scope: NamespaceBinding, child: Node*): Elem =
     scala.xml.Elem.apply(prefix, label, attributes, scope, minimizeEmpty = true, child: _*)
 
-  lazy val parsedxml1: Elem = XML.load(new InputSource(new StringReader("<hello><world/></hello>")))
-  lazy val parsedxml11: Elem = XML.load(new InputSource(new StringReader("<hello><world/></hello>")))
+  lazy val parsedxml1: Elem = XML.loadString("<hello><world/></hello>")
+  lazy val parsedxml11: Elem = XML.loadString("<hello><world/></hello>")
   val xmlFile2: String = "<bib><book><author>Peter Buneman</author><author>Dan Suciu</author><title>Data on ze web</title></book><book><author>John Mitchell</author><title>Foundations of Programming Languages</title></book></bib>"
-  lazy val parsedxml2: Elem = XML.load(new InputSource(new StringReader(xmlFile2)))
+  lazy val parsedxml2: Elem = XML.loadString(xmlFile2)
 
   @UnitTest
   def equality(): Unit = {
@@ -46,9 +46,7 @@ class XMLTestJVM {
     assertTrue(Array(parsedxml1).toList sameElements List(parsedxml11))
 
     val x2: String = "<book><author>Peter Buneman</author><author>Dan Suciu</author><title>Data on ze web</title></book>"
-
-    val i: InputSource = new InputSource(new StringReader(x2))
-    val x2p: Elem = XML.load(i)
+    val x2p: Elem = XML.loadString(x2)
 
     assertEquals(Elem(null, "book", e, sc,
       Elem(null, "author", e, sc, Text("Peter Buneman")),
@@ -146,8 +144,7 @@ class XMLTestJVM {
     val xmlAttrValueNorm: String = "<personne id='p0003' nom='&#x015e;ahingšz' />"
 
     {
-      val isrcA: InputSource = new InputSource(new StringReader(xmlAttrValueNorm))
-      val parsedxmlA: Elem = XML.load(isrcA)
+      val parsedxmlA: Elem = XML.loadString(xmlAttrValueNorm)
       val c: Char = (parsedxmlA \ "@nom").text.charAt(0)
       assertTrue(c == '\u015e')
     }
@@ -689,6 +686,17 @@ class XMLTestJVM {
     assertTrue(gotAnError)
   }
 
+  // Here we see that opening InputStream prematurely, as was done previously, breaks XInclude.
+  @UnitTest(expected = classOf[org.xml.sax.SAXParseException]) def xIncludeNeedsSystemId(): Unit = {
+    val parserFactory = xercesInternal
+    parserFactory.setNamespaceAware(true)
+    parserFactory.setXIncludeAware(true)
+    XML
+      .withSAXParser(parserFactory.newSAXParser)
+      .load(getClass.getResource("site.xml").openStream())
+      .toString
+  }
+
   // Now that we can use XML parser configured to be namespace-aware,
   // we can also configure it to be XInclude-aware and process XML Includes:
   def check(
@@ -700,7 +708,7 @@ class XMLTestJVM {
     parserFactory.setXIncludeAware(true)
     val actual: String = XML
       .withSAXParser(parserFactory.newSAXParser)
-      .load(getClass.getResource(resourceName).toString)
+      .load(getClass.getResource(resourceName))
       .toString
 
     assertEquals(expected, actual)

shared/src/main/scala/scala/xml/XML.scala

@@ -19,13 +19,13 @@ import java.nio.channels.Channels
 import scala.util.control.Exception.ultimately
 
 object Source {
-  def fromFile(file: File): InputSource = new InputSource(new FileInputStream(file))
-  def fromFile(fd: FileDescriptor): InputSource = new InputSource(new FileInputStream(fd))
-  def fromFile(name: String): InputSource = new InputSource(new FileInputStream(name))
-
+  def fromFile(name: String): InputSource = fromFile(new File(name))
+  def fromFile(file: File): InputSource = fromUrl(file.toURI.toURL)
+  def fromUrl(url: java.net.URL): InputSource = fromSysId(url.toString)
+  def fromSysId(sysID: String): InputSource = new InputSource(sysID)
+  def fromFile(fd: FileDescriptor): InputSource = fromInputStream(new FileInputStream(fd))
   def fromInputStream(is: InputStream): InputSource = new InputSource(is)
   def fromReader(reader: Reader): InputSource = new InputSource(reader)
-  def fromSysId(sysID: String): InputSource = new InputSource(sysID)
   def fromString(string: String): InputSource = fromReader(new StringReader(string))
 }
 

shared/src/main/scala/scala/xml/factory/XMLLoader.scala

@@ -14,32 +14,36 @@ package scala
 package xml
 package factory
 
-import org.xml.sax.{SAXNotRecognizedException, XMLReader}
+import org.xml.sax.{SAXNotRecognizedException, SAXNotSupportedException, XMLReader}
 import javax.xml.parsers.SAXParserFactory
 import parsing.{FactoryAdapter, NoBindingFactoryAdapter}
 import java.io.{File, FileDescriptor, InputStream, Reader}
 import java.net.URL
 
 /**
  * Presents collection of XML loading methods which use the parser
- *  created by "def parser".
+ *  created by "def parser" or the reader created by "def reader".
  */
 trait XMLLoader[T <: Node] {
   import scala.xml.Source._
   def adapter: FactoryAdapter = new NoBindingFactoryAdapter()
 
+  private def setSafeDefaults(parserFactory: SAXParserFactory): Unit = {
+    parserFactory.setFeature("http://javax.xml.XMLConstants/feature/secure-processing", true)
+    parserFactory.setFeature("http://apache.org/xml/features/nonvalidating/load-external-dtd", false)
+    parserFactory.setFeature("http://apache.org/xml/features/disallow-doctype-decl", true)
+    parserFactory.setFeature("http://xml.org/sax/features/external-parameter-entities", false)
+    parserFactory.setFeature("http://xml.org/sax/features/external-general-entities", false)
+    parserFactory.setFeature("http://xml.org/sax/features/resolve-dtd-uris", false)
+    parserFactory.setXIncludeAware(false)
+    parserFactory.setNamespaceAware(false)
+  }
+
   private lazy val parserInstance: ThreadLocal[SAXParser] = new ThreadLocal[SAXParser] {
     override def initialValue: SAXParser = {
-      val parser: SAXParserFactory = SAXParserFactory.newInstance
-      parser.setFeature("http://javax.xml.XMLConstants/feature/secure-processing", true)
-      parser.setFeature("http://apache.org/xml/features/nonvalidating/load-external-dtd", false)
-      parser.setFeature("http://apache.org/xml/features/disallow-doctype-decl", true)
-      parser.setFeature("http://xml.org/sax/features/external-parameter-entities", false)
-      parser.setFeature("http://xml.org/sax/features/external-general-entities", false)
-      parser.setFeature("http://xml.org/sax/features/resolve-dtd-uris", false)
-      parser.setXIncludeAware(false)
-      parser.setNamespaceAware(false)
-      parser.newSAXParser
+      val parserFactory: SAXParserFactory = SAXParserFactory.newInstance
+      setSafeDefaults(parserFactory)
+      parserFactory.newSAXParser
     }
   }
 
@@ -51,72 +55,68 @@ trait XMLLoader[T <: Node] {
 
   /**
    * Loads XML from the given InputSource, using the supplied parser.
-   *  The methods available in scala.xml.XML use the XML parser in the JDK.
+   * The methods available in scala.xml.XML use the XML parser in the JDK
+   * (unless another parser is present on the classpath).
    */
-  def loadXML(source: InputSource, parser: SAXParser): T = loadXML(source, parser.getXMLReader)
+  def loadXML(inputSource: InputSource, parser: SAXParser): T = loadXML(inputSource, parser.getXMLReader)
 
-  def loadXMLNodes(source: InputSource, parser: SAXParser): Seq[Node] = loadXMLNodes(source, parser.getXMLReader)
+  def loadXMLNodes(inputSource: InputSource, parser: SAXParser): Seq[Node] = loadXMLNodes(inputSource, parser.getXMLReader)
 
-  private def loadXML(source: InputSource, reader: XMLReader): T = {
-    val result: FactoryAdapter = parse(source, reader)
+  private def loadXML(inputSource: InputSource, reader: XMLReader): T = {
+    val result: FactoryAdapter = parse(inputSource, reader)
     result.rootElem.asInstanceOf[T]
   }
-  
-  private def loadXMLNodes(source: InputSource, reader: XMLReader): Seq[Node] = {
-    val result: FactoryAdapter = parse(source, reader)
+
+  private def loadXMLNodes(inputSource: InputSource, reader: XMLReader): Seq[Node] = {
+    val result: FactoryAdapter = parse(inputSource, reader)
     result.prolog ++ (result.rootElem :: result.epilogue)
   }
 
-  private def parse(source: InputSource, reader: XMLReader): FactoryAdapter = {
-    if (source == null) throw new IllegalArgumentException("InputSource cannot be null")
+  private def parse(inputSource: InputSource, xmlReader: XMLReader): FactoryAdapter = {
+    if (inputSource == null) throw new IllegalArgumentException("InputSource cannot be null")
 
     val result: FactoryAdapter = adapter
 
-    reader.setContentHandler(result)
-    reader.setDTDHandler(result)
+    xmlReader.setContentHandler(result)
+    xmlReader.setDTDHandler(result)
     /* Do not overwrite pre-configured EntityResolver. */
-    if (reader.getEntityResolver == null) reader.setEntityResolver(result)
+    if (xmlReader.getEntityResolver == null) xmlReader.setEntityResolver(result)
     /* Do not overwrite pre-configured ErrorHandler. */
-    if (reader.getErrorHandler == null) reader.setErrorHandler(result)
+    if (xmlReader.getErrorHandler == null) xmlReader.setErrorHandler(result)
 
     try {
-      reader.setProperty("http://xml.org/sax/properties/lexical-handler", result)
+      xmlReader.setProperty("http://xml.org/sax/properties/lexical-handler", result)
     } catch {
       case _: SAXNotRecognizedException =>
+      case _: SAXNotSupportedException =>
     }
 
     result.scopeStack = TopScope :: result.scopeStack
-    reader.parse(source)
+    xmlReader.parse(inputSource)
     result.scopeStack = result.scopeStack.tail
 
     result
   }
 
-  /** loads XML from given InputSource. */
-  def load(source: InputSource): T = loadXML(source, reader)
-
-  /** Loads XML from the given file, file descriptor, or filename. */
+  /** Loads XML. */
+  def load(inputSource: InputSource): T = loadXML(inputSource, reader)
+  def loadFile(fileName: String): T = load(fromFile(fileName))
   def loadFile(file: File): T = load(fromFile(file))
-  def loadFile(fd: FileDescriptor): T = load(fromFile(fd))
-  def loadFile(name: String): T = load(fromFile(name))
-
-  /** loads XML from given InputStream, Reader, sysID, or URL. */
-  def load(is: InputStream): T = load(fromInputStream(is))
+  def load(url: URL): T = load(fromUrl(url))
+  def load(sysId: String): T = load(fromSysId(sysId))
+  def loadFile(fileDescriptor: FileDescriptor): T = load(fromFile(fileDescriptor))
+  def load(inputStream: InputStream): T = load(fromInputStream(inputStream))
   def load(reader: Reader): T = load(fromReader(reader))
-  def load(sysID: String): T = load(fromSysId(sysID))
-  def load(url: URL): T = load(fromInputStream(url.openStream()))
-
-  /** Loads XML from the given String. */
   def loadString(string: String): T = load(fromString(string))
 
   /** Load XML nodes, including comments and processing instructions that precede and follow the root element. */
-  def loadNodes(source: InputSource): Seq[Node] = loadXMLNodes(source, reader)
+  def loadNodes(inputSource: InputSource): Seq[Node] = loadXMLNodes(inputSource, reader)
+  def loadFileNodes(fileName: String): Seq[Node] = loadNodes(fromFile(fileName))
   def loadFileNodes(file: File): Seq[Node] = loadNodes(fromFile(file))
-  def loadFileNodes(fd: FileDescriptor): Seq[Node] = loadNodes(fromFile(fd))
-  def loadFileNodes(name: String): Seq[Node] = loadNodes(fromFile(name))
-  def loadNodes(is: InputStream): Seq[Node] = loadNodes(fromInputStream(is))
+  def loadNodes(url: URL): Seq[Node] = loadNodes(fromUrl(url))
+  def loadNodes(sysId: String): Seq[Node] = loadNodes(fromSysId(sysId))
+  def loadFileNodes(fileDescriptor: FileDescriptor): Seq[Node] = loadNodes(fromFile(fileDescriptor))
+  def loadNodes(inputStream: InputStream): Seq[Node] = loadNodes(fromInputStream(inputStream))
   def loadNodes(reader: Reader): Seq[Node] = loadNodes(fromReader(reader))
-  def loadNodes(sysID: String): Seq[Node] = loadNodes(fromSysId(sysID))
-  def loadNodes(url: URL): Seq[Node] = loadNodes(fromInputStream(url.openStream()))
   def loadStringNodes(string: String): Seq[Node] = loadNodes(fromString(string))
 }