nokogiri 1.5.0.beta.3 → 1.5.0.beta.4

data/ext/java/nokogiri/XmlText.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -42,7 +42,6 @@ import org.jruby.Ruby;
 import org.jruby.RubyClass;
 import org.jruby.anno.JRubyClass;
 import org.jruby.anno.JRubyMethod;
-import org.jruby.javasupport.JavaUtil;
 import org.jruby.runtime.ThreadContext;
 import org.jruby.runtime.builtin.IRubyObject;
 import org.w3c.dom.Document;
@@ -51,6 +50,8 @@ import org.w3c.dom.Node;
 /**
  * Class for Nokogiri::XML::Text
  *
+ * @author sergio
+ * @author Yoko Harada <yokolet@gmail.com>
  */
 @JRubyClass(name="Nokogiri::XML::Text", parent="Nokogiri::XML::CharacterData")
 public class XmlText extends XmlNode {

data/ext/java/nokogiri/XmlXpathContext.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -32,6 +32,8 @@
 
 package nokogiri;
 
+import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
+
 import java.util.Set;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
@@ -63,6 +65,8 @@ import org.w3c.dom.NodeList;
 /**
  * Class for Nokogiri::XML::XpathContext
  *
+ * @author sergio
+ * @author Yoko Harada <yokolet@gmail.com>
  */
 @JRubyClass(name="Nokogiri::XML::XPathContext")
 public class XmlXpathContext extends RubyObject {
@@ -98,7 +102,6 @@ public class XmlXpathContext extends RubyObject {
             }
             XPathExpression xpathExpression = xpath.compile(src);
             return node_set(context, xpathExpression);
-            //return new XmlXpath(context.getRuntime(), (RubyClass)context.getRuntime().getClassFromPath("Nokogiri::XML::XPath"), xpathExpression, this.context);
         } catch (XPathExpressionException xpee) {
             xpee = new XPathExpressionException(src);
             RubyException e =
@@ -111,7 +114,6 @@ public class XmlXpathContext extends RubyObject {
         XmlNodeSet result = null;
         try {  
             result = tryGetNodeSet(xpathExpression);
-//            result.relink_namespace(context);
             result.setDocument(context.document(rbctx));
             return result;
         } catch (XPathExpressionException xpee) {
@@ -126,7 +128,9 @@ public class XmlXpathContext extends RubyObject {
     
     private XmlNodeSet tryGetNodeSet(XPathExpression xpathExpression) throws XPathExpressionException {
         NodeList nodes = (NodeList)xpathExpression.evaluate(context.node, XPathConstants.NODESET);
-        return new XmlNodeSet(getRuntime(), nodes);       
+        XmlNodeSet xmlNodeSet = (XmlNodeSet) NokogiriService.XML_NODESET_ALLOCATOR.allocate(getRuntime(), getNokogiriClass(getRuntime(), "Nokogiri::XML::NodeSet"));
+        xmlNodeSet.setNodeList(nodes);
+        return xmlNodeSet;    
     }
     
     private static Pattern number_pattern = Pattern.compile("\\d.*");

data/ext/java/nokogiri/XsltStylesheet.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -61,6 +61,8 @@ import org.w3c.dom.Document;
 /**
  * Class for Nokogiri::XSLT::Stylesheet
  *
+ * @author sergio
+ * @author Yoko Harada <yokolet@gmail.com>
  */
 @JRubyClass(name="Nokogiri::XSLT::Stylesheet")
 public class XsltStylesheet extends RubyObject {
@@ -140,7 +142,7 @@ public class XsltStylesheet extends RubyObject {
 
     @JRubyMethod(rest = true, required=1, optional=2)
     public IRubyObject transform(ThreadContext context, IRubyObject[] args) {
-        Ruby ruby = context.getRuntime();
+        Ruby runtime = context.getRuntime();
 
         DOMSource docSource = new DOMSource(((XmlDocument) args[0]).getDocument());
         DOMResult result = new DOMResult();
@@ -152,19 +154,19 @@ public class XsltStylesheet extends RubyObject {
             }
             transf.transform(docSource, result);
         } catch(TransformerConfigurationException ex) {
-            throw ruby.newRuntimeError("Could not transform the document.");
+            throw runtime.newRuntimeError("Could not transform the document.");
         } catch(TransformerException ex) {
-            throw ruby.newRuntimeError("Could not transform the document.");
+            throw runtime.newRuntimeError("Could not transform the document.");
         }
         
         if ("html".equals(result.getNode().getFirstChild().getNodeName())) {
-            return new HtmlDocument(ruby,
-                    getNokogiriClass(ruby, "Nokogiri::HTML::Document"),
-                    (Document) result.getNode());
+            HtmlDocument htmlDocument = (HtmlDocument) getNokogiriClass(runtime, "Nokogiri::HTML::Document").allocate();
+            htmlDocument.setNode(context, (Document) result.getNode());
+            return htmlDocument;
         } else {
-            return new XmlDocument(ruby,
-                    getNokogiriClass(ruby, "Nokogiri::XML::Document"),
-                    (Document) result.getNode());
+            XmlDocument xmlDocument = (XmlDocument) NokogiriService.XML_DOCUMENT_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::Document"));
+            xmlDocument.setNode(context, (Document) result.getNode());
+            return xmlDocument;
         }
     }
     

data/ext/java/nokogiri/internals/HtmlDomParserContext.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -35,6 +35,7 @@ package nokogiri.internals;
 import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
 import static nokogiri.internals.NokogiriHelpers.isNamespace;
 import nokogiri.HtmlDocument;
+import nokogiri.NokogiriService;
 import nokogiri.XmlDocument;
 
 import org.apache.xerces.parsers.DOMParser;
@@ -56,6 +57,7 @@ import org.w3c.dom.Document;
  * Parser for HtmlDocument. This class actually parses HtmlDocument using NekoHtml.
  * 
  * @author sergio
+ * @author Yoko Harada <yokolet@gmail.com>
  */
 public class HtmlDomParserContext extends XmlDomParserContext {
 
@@ -108,16 +110,15 @@ public class HtmlDomParserContext extends XmlDomParserContext {
     @Override
     protected XmlDocument getNewEmptyDocument(ThreadContext context) {
         IRubyObject[] args = new IRubyObject[0];
-        return (XmlDocument) XmlDocument.rbNew(context,
-                    getNokogiriClass(context.getRuntime(), "Nokogiri::HTML::Document"),
-                    args);
+        return (XmlDocument) XmlDocument.rbNew(context, getNokogiriClass(context.getRuntime(), "Nokogiri::HTML::Document"), args);
     }
 
     @Override
     protected XmlDocument wrapDocument(ThreadContext context,
-                                       RubyClass klass,
-                                       Document doc) {
-        HtmlDocument htmlDocument = new HtmlDocument(context.getRuntime(), klass, doc);
+                                       RubyClass klazz,
+                                       Document document) {
+        HtmlDocument htmlDocument = (HtmlDocument) NokogiriService.HTML_DOCUMENT_ALLOCATOR.allocate(context.getRuntime(), klazz);
+        htmlDocument.setNode(context, document);
         htmlDocument.setEncoding(ruby_encoding);
         return htmlDocument;
     }

data/ext/java/nokogiri/internals/NokogiriDocumentCache.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]

data/ext/java/nokogiri/internals/NokogiriErrorHandler.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -32,9 +32,12 @@
 
 package nokogiri.internals;
 
+import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
+
 import java.util.ArrayList;
 import java.util.List;
 
+import nokogiri.NokogiriService;
 import nokogiri.XmlSyntaxError;
 
 import org.apache.xerces.xni.parser.XMLErrorHandler;
@@ -50,6 +53,7 @@ import org.xml.sax.ErrorHandler;
  * uses this type of the error handler.
  * 
  * @author sergio
+ * @author Yoko Harada <yokolet@gmail.com>
  */
 public abstract class NokogiriErrorHandler implements ErrorHandler, XMLErrorHandler {
     protected List<Exception> errors;
@@ -64,11 +68,13 @@ public abstract class NokogiriErrorHandler implements ErrorHandler, XMLErrorHand
 
     public List<Exception> getErrors() { return errors; }
 
-    public List<IRubyObject> getErrorsReadyForRuby(ThreadContext context){
-        Ruby ruby = context.getRuntime();
+    public List<IRubyObject> getErrorsReadyForRuby(ThreadContext context) {
+        Ruby runtime = context.getRuntime();
         List<IRubyObject> res = new ArrayList<IRubyObject>();
-        for(int i = 0; i < errors.size(); i++) {
-            res.add(new XmlSyntaxError(ruby, errors.get(i)));
+        for (int i = 0; i < errors.size(); i++) {
+            XmlSyntaxError xmlSyntaxError = (XmlSyntaxError) NokogiriService.XML_SYNTAXERROR_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::SyntaxError"));
+            xmlSyntaxError.setException(errors.get(i));
+            res.add(xmlSyntaxError);
         }
         return res;
     }

data/ext/java/nokogiri/internals/NokogiriHandler.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -59,6 +59,7 @@ import org.xml.sax.ext.DefaultHandler2;
  * A handler for SAX parsing.
  * 
  * @author sergio
+ * @author Yoko Harada <yokolet@gmail.com>
  */
 public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
 
@@ -79,13 +80,17 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
     private Locator locator;
     private ArrayDeque<Integer> lines;
     private ArrayDeque<Integer> columns;
+    private static String htmlParserName = "Nokogiri::HTML::SAX::Parser";
+    private boolean needEmptyAttrCheck = false;
 
-    public NokogiriHandler(Ruby ruby, IRubyObject object) {
-        this.ruby = ruby;
-        this.attrClass = (RubyClass) ruby.getClassFromPath("Nokogiri::XML::SAX::Parser::Attribute");
+    public NokogiriHandler(Ruby runtime, IRubyObject object) {
+        this.ruby = runtime;
+        this.attrClass = (RubyClass) runtime.getClassFromPath("Nokogiri::XML::SAX::Parser::Attribute");
         this.object = object;
         lines = new ArrayDeque<Integer>();
         columns = new ArrayDeque<Integer>();
+        String objectName = object.getMetaClass().getName();
+        if (htmlParserName.equals(objectName)) needEmptyAttrCheck = true;
     }
     
     public void setDocumentLocator(Locator locator) {
@@ -150,11 +155,22 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
                 ns.add(ruby.newString(val));
                 rubyNSAttr.add(ns);
             } else {
-                IRubyObject[] args = new IRubyObject[4];
-                args[0] = stringOrNil(ruby, ln);
-                args[1] = stringOrNil(ruby, pre);
-                args[2] = stringOrNil(ruby, u);
-                args[3] = stringOrNil(ruby, val);
+                IRubyObject[] args = null;
+                if (needEmptyAttrCheck) {
+                    if (isEmptyAttr(ln)) {
+                        args = new IRubyObject[3];
+                        args[0] = stringOrNil(ruby, ln);
+                        args[1] = stringOrNil(ruby, pre);
+                        args[2] = stringOrNil(ruby, u);
+                    }
+                } 
+                if (args == null) {
+                    args = new IRubyObject[4];
+                    args[0] = stringOrNil(ruby, ln);
+                    args[1] = stringOrNil(ruby, pre);
+                    args[2] = stringOrNil(ruby, u);
+                    args[3] = stringOrNil(ruby, val);
+                }
 
                 IRubyObject attr = RuntimeHelpers.invoke(context, attrClass, "new", args);
                 rubyAttr.add(attr);
@@ -170,6 +186,17 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
              rubyNSAttr);
     }
     
+    private static String[] emptyAttrs =
+        {"checked", "compact", "declare", "defer", "disabled", "ismap", "multiple", 
+         "noresize", "nohref", "noshade", "nowrap", "readonly", "selected"};
+    
+    private boolean isEmptyAttr(String name) {
+        for (String emptyAttr : emptyAttrs) {
+            if (emptyAttr.equals(name)) return true;
+        }
+        return false;
+    }
+    
     public Integer getLine() {
         return lines.pop();
     }

data/ext/java/nokogiri/internals/NokogiriHelpers.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -50,7 +50,6 @@ import org.jruby.Ruby;
 import org.jruby.RubyArray;
 import org.jruby.RubyClass;
 import org.jruby.RubyEncoding;
-import org.jruby.RubyHash;
 import org.jruby.RubyString;
 import org.jruby.runtime.builtin.IRubyObject;
 import org.jruby.util.ByteList;
@@ -63,6 +62,8 @@ import org.w3c.dom.NodeList;
  * A class for various utility methods.
  * 
  * @author serabe
+ * @author Patrick Mahoney <pat@polycrystal.org>
+ * @author Yoko Harada <yokolet@gmail.com>
  */
 public class NokogiriHelpers {
     public static final String CACHED_NODE = "NOKOGIRI_CACHED_NODE";
@@ -103,47 +104,45 @@ public class NokogiriHelpers {
      * subclass of XmlNode is chosen based on the type of
      * <code>node</code>.
      */
-    public static IRubyObject constructNode(Ruby ruby, Node node) {
-        if (node == null) return ruby.getNil();
+    public static IRubyObject constructNode(Ruby runtime, Node node) {
+        if (node == null) return runtime.getNil();
         // this is slow; need a way to cache nokogiri classes/modules somewhere
         switch (node.getNodeType()) {
             case Node.ELEMENT_NODE:
-                XmlElement xmlElement = (XmlElement) getNokogiriClass(ruby, "Nokogiri::XML::Element").allocate();
-                xmlElement.setNode(ruby.getCurrentContext(), node);
+                XmlElement xmlElement = (XmlElement) NokogiriService.XML_ELEMENT_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::Element"));
+                xmlElement.setNode(runtime.getCurrentContext(), node);
                 return xmlElement;
             case Node.ATTRIBUTE_NODE:
-                XmlAttr xmlAttr = (XmlAttr) getNokogiriClass(ruby, "Nokogiri::XML::Attr").allocate();
-                xmlAttr.setNode(ruby.getCurrentContext(), node);
+                XmlAttr xmlAttr = (XmlAttr) NokogiriService.XML_ATTR_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::Attr"));
+                xmlAttr.setNode(runtime.getCurrentContext(), node);
                 return xmlAttr;
             case Node.TEXT_NODE:
-                XmlText xmlText = (XmlText) getNokogiriClass(ruby, "Nokogiri::XML::Text").allocate();
-                xmlText.setNode(ruby.getCurrentContext(), node);
+                XmlText xmlText = (XmlText) NokogiriService.XML_TEXT_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::Text"));
+                xmlText.setNode(runtime.getCurrentContext(), node);
                 return xmlText;
             case Node.COMMENT_NODE:
-                XmlComment xmlComment = (XmlComment) getNokogiriClass(ruby, "Nokogiri::XML::Comment").allocate();
-                xmlComment.setNode(ruby.getCurrentContext(), node);
+                XmlComment xmlComment = (XmlComment) NokogiriService.XML_COMMENT_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::Comment"));
+                xmlComment.setNode(runtime.getCurrentContext(), node);
                 return xmlComment;
             case Node.ENTITY_NODE:
-                return new XmlNode(ruby, getNokogiriClass(ruby, "Nokogiri::XML::EntityDecl"), node);
+                return new XmlNode(runtime, getNokogiriClass(runtime, "Nokogiri::XML::EntityDecl"), node);
             case Node.CDATA_SECTION_NODE:
-                XmlCdata xmlCdata = (XmlCdata) getNokogiriClass(ruby, "Nokogiri::XML::CDATA").allocate();
-                xmlCdata.setNode(ruby.getCurrentContext(), node);
+                XmlCdata xmlCdata = (XmlCdata) NokogiriService.XML_CDATA_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::CDATA"));
+                xmlCdata.setNode(runtime.getCurrentContext(), node);
                 return xmlCdata;
             case Node.DOCUMENT_NODE:
-                XmlDocument xmlDocument = (XmlDocument) getNokogiriClass(ruby, "Nokogiri::XML::Document").allocate();
-                xmlDocument.setNode(ruby.getCurrentContext(), node);
+                XmlDocument xmlDocument = (XmlDocument) NokogiriService.XML_DOCUMENT_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::Document"));
+                xmlDocument.setNode(runtime.getCurrentContext(), node);
                 return xmlDocument;
             default:
-                XmlNode xmlNode = (XmlNode) getNokogiriClass(ruby, "Nokogiri::XML::Node").allocate();
-                xmlNode.setNode(ruby.getCurrentContext(), node);
+                XmlNode xmlNode = (XmlNode) NokogiriService.XML_NODE_ALLOCATOR.allocate(runtime, getNokogiriClass(runtime, "Nokogiri::XML::Node"));
+                xmlNode.setNode(runtime.getCurrentContext(), node);
                 return xmlNode;
         }
     }
     
     public static RubyClass getNokogiriClass(Ruby ruby, String name) {
-        RubyHash classCache = (RubyHash) ruby.getGlobalVariables().get(NokogiriService.nokogiriClassCacheGvarName);
-        IRubyObject rubyName = RubyString.newString(ruby, name);
-        return (RubyClass)classCache.fastARef(rubyName);
+        return NokogiriService.nokogiriClassCache.get(name);
     }
 
     public static IRubyObject stringOrNil(Ruby runtime, String s) {

data/ext/java/nokogiri/internals/NokogiriNamespaceCache.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]
@@ -36,10 +36,11 @@ import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
 import static nokogiri.internals.NokogiriHelpers.isNamespace;
 
 import java.util.ArrayList;
-import java.util.HashMap;
+import java.util.LinkedHashMap;
 import java.util.List;
 import java.util.Map;
 
+import nokogiri.NokogiriService;
 import nokogiri.XmlDocument;
 import nokogiri.XmlNamespace;
 
@@ -62,7 +63,7 @@ public class NokogiriNamespaceCache {
 
     public NokogiriNamespaceCache() {
         keys = new ArrayList<Long>();
-        cache = new HashMap<Integer, CacheEntry>();
+        cache = new LinkedHashMap<Integer, CacheEntry>();
     }
     
     private Long hashCode(String prefix, String href) {
@@ -121,7 +122,7 @@ public class NokogiriNamespaceCache {
             keys.add(hash);
             index = keys.size() - 1;
             String actualPrefix = (prefix.equals("")) ? null : prefix;
-            XmlNamespace namespace = (XmlNamespace) getNokogiriClass(ruby, "Nokogiri::XML::Namespace").allocate();
+            XmlNamespace namespace = (XmlNamespace) NokogiriService.XML_NAMESPACE_ALLOCATOR.allocate(ruby, getNokogiriClass(ruby, "Nokogiri::XML::Namespace"));
             namespace.setDefinition(ruby, actualPrefix, href);
             namespace.setDocument(document);
             CacheEntry entry = new CacheEntry(namespace, node);

data/ext/java/nokogiri/internals/NokogiriNamespaceContext.java

@@ -1,7 +1,7 @@
 /**
  * (The MIT License)
  *
- * Copyright (c) 2008 - 2010:
+ * Copyright (c) 2008 - 2011:
  *
  * * {Aaron Patterson}[http://tenderlovemaking.com]
  * * {Mike Dalessio}[http://mike.daless.io]