nokogiri 1.5.6.rc1-java → 1.5.6.rc2-java

data/CHANGELOG.ja.rdoc

@@ -6,6 +6,9 @@
 
   * JRuby で '#' で始まる文字列を名前とする EntityReference を作ろうとすると INVALID_CHARACTER_ERR という例外がはっせいする。 #719
   * JRuby では Nodeのサブクラスのnamespaceを正しく文字列に変換しない。 #715
+  * Nokogiri now detects XSLT transform errors. #731 (Thanks, Justin Fitzsimmons!)
+  * Raise an ArgumentError if an invalid encoding is passed to the SAX parser. #756 (Thanks, Bradley Schaefer!)
+  * JRuby Node#content now renders newlines properly. #737 (Thanks, Piotr Szmielew!)
 
 
 == 1.5.5 / ２０１２年６月２４日

data/CHANGELOG.rdoc

@@ -6,6 +6,9 @@
 
   * JRuby raises INVALID_CHARACTER_ERR exception when EntityReference name starts with '#'. #719
   * JRuby doesn't coerce namespaces out of strings on a direct subclass of Node. #715
+  * Nokogiri now detects XSLT transform errors. #731 (Thanks, Justin Fitzsimmons!)
+  * Raise an ArgumentError if an invalid encoding is passed to the SAX parser. #756 (Thanks, Bradley Schaefer!)
+  * JRuby Node#content now renders newlines properly. #737 (Thanks, Piotr Szmielew!)
 
 
 == 1.5.5 / 2012-06-24

data/Manifest.txt

@@ -43,8 +43,13 @@ ext/java/nokogiri/XmlSyntaxError.java
 ext/java/nokogiri/XmlText.java
 ext/java/nokogiri/XmlXpathContext.java
 ext/java/nokogiri/XsltStylesheet.java
+ext/java/nokogiri/internals/ClosedStreamException.java
 ext/java/nokogiri/internals/HtmlDomParserContext.java
+ext/java/nokogiri/internals/NokogiriBlockingQueueInputStream.java
 ext/java/nokogiri/internals/NokogiriDocumentCache.java
+ext/java/nokogiri/internals/NokogiriDomParser.java
+ext/java/nokogiri/internals/NokogiriEncodingReaderWrapper.java
+ext/java/nokogiri/internals/NokogiriEntityResolver.java
 ext/java/nokogiri/internals/NokogiriErrorHandler.java
 ext/java/nokogiri/internals/NokogiriHandler.java
 ext/java/nokogiri/internals/NokogiriHelpers.java
@@ -62,7 +67,6 @@ ext/java/nokogiri/internals/ReaderNode.java
 ext/java/nokogiri/internals/SaveContextVisitor.java
 ext/java/nokogiri/internals/SchemaErrorHandler.java
 ext/java/nokogiri/internals/XmlDeclHandler.java
-ext/java/nokogiri/internals/XmlDomParser.java
 ext/java/nokogiri/internals/XmlDomParserContext.java
 ext/java/nokogiri/internals/XmlSaxParser.java
 ext/java/nokogiri/internals/XsltExtensionFunction.java
@@ -230,13 +234,13 @@ test/files/snuggles.xml
 test/files/staff.dtd
 test/files/staff.xml
 test/files/staff.xslt
+test/files/test_document_url/bar.xml
+test/files/test_document_url/document.dtd
+test/files/test_document_url/document.xml
 test/files/tlm.html
 test/files/to_be_xincluded.xml
 test/files/valid_bar.xml
 test/files/xinclude.xml
-test/files/test_document_url/bar.xml
-test/files/test_document_url/document.dtd
-test/files/test_document_url/document.xml
 test/helper.rb
 test/html/sax/test_parser.rb
 test/html/sax/test_parser_context.rb

data/README.ja.rdoc

@@ -1,4 +1,4 @@
-= Nokogiri (鋸)
+= Nokogiri (鋸) {<img src="https://secure.travis-ci.org/sparklemotion/nokogiri.png?rvm=1.9.3" />}[http://travis-ci.org/sparklemotion/nokogiri] {<img src="https://codeclimate.com/badge.png" />}[https://codeclimate.com/github/sparklemotion/nokogiri]
 
 * http://nokogiri.org/
 * http://github.com/sparklemotion/nokogiri/wikis

data/README.rdoc

@@ -1,4 +1,4 @@
-= Nokogiri {<img src="https://secure.travis-ci.org/sparklemotion/nokogiri.png?rvm=1.9.3" />}[http://travis-ci.org/sparklemotion/nokogiri]
+= Nokogiri {<img src="https://secure.travis-ci.org/sparklemotion/nokogiri.png?rvm=1.9.3" />}[http://travis-ci.org/sparklemotion/nokogiri] {<img src="https://codeclimate.com/badge.png" />}[https://codeclimate.com/github/sparklemotion/nokogiri]
 
 * http://nokogiri.org
 * http://github.com/sparklemotion/nokogiri/wikis

data/ROADMAP.md

@@ -69,6 +69,9 @@
   * we should standardize on a hash of options for these and other calls
 * what should NodeSet#xpath return?
   * https://github.com/sparklemotion/nokogiri/issues/656
+* also, clean up or unify the implementations of #xpath-and-friends in Node and NodeSet
+  * implementations are very similar, but no shared code :(
+  * decorate nodes in a consistent manner
 
 ## Encoding
 

data/Rakefile

@@ -46,14 +46,10 @@ HOE = Hoe.spec 'nokogiri' do
     ["mini_portile",    ">= 0.2.2"],
     ["minitest",        "~> 2.2.2"],
     ["rake",            ">= 0.9"],
-    ["rake-compiler",   "=  0.8.0"]
+    ["rake-compiler",   "=  0.8.0"],
+    ["racc",            ">= 1.4.6"],
+    ["rexical",         ">= 1.0.5"]
   ]
-  if ! java?
-    self.extra_dev_deps += [
-      ["racc",            ">= 1.4.6"],
-      ["rexical",         ">= 1.0.5"]
-    ]
-  end
 
   if java?
     self.spec_extras = { :platform => 'java' }
@@ -118,6 +114,20 @@ desc "Generate css/parser.rb and css/tokenizer.rex"
 task 'generate' => [GENERATED_PARSER, GENERATED_TOKENIZER]
 task 'gem:spec' => 'generate' if Rake::Task.task_defined?("gem:spec")
 
+# This is a big hack to make sure that the racc and rexical
+# dependencies in the Gemfile are constrainted to ruby platforms
+# (i.e. MRI and Rubinius). There's no way to do that through hoe,
+# and any solution will require changing hoe and hoe-bundler.
+old_gemfile_task = Rake::Task['bundler:gemfile']
+task 'bundler:gemfile' do
+  old_gemfile_task.invoke
+
+  lines = File.open('Gemfile', 'r') { |f| f.readlines }.map do |line|
+    line =~ /racc|rexical/ ? "#{line.strip}, :platform => :ruby" : line
+  end
+  File.open('Gemfile', 'w') { |f| lines.each { |line| f.puts line } }
+end
+
 file GENERATED_PARSER => "lib/nokogiri/css/parser.y" do |t|
   racc = RbConfig::CONFIG['target_os'] =~ /mswin32/ ? '' : `which racc`.strip
   racc = "#{::RbConfig::CONFIG['bindir']}/racc" if racc.empty?
@@ -149,9 +159,18 @@ task :java_debug do
   ENV['JAVA_OPTS'] = '-Xdebug -Xrunjdwp:transport=dt_socket,address=8000,server=y,suspend=y' if java? && ENV['JAVA_DEBUG']
 end
 
+if java?
+  task :test_18 => :test
+  task :test_19 do
+    ENV['JRUBY_OPTS'] = "--1.9"
+    Rake::Task["test"].invoke
+  end
+end
+
 Rake::Task[:test].prerequisites << :compile
 Rake::Task[:test].prerequisites << :java_debug
 Rake::Task[:test].prerequisites << :check_extra_deps unless java?
+
 if Hoe.plugins.include?(:debugging)
   ['valgrind', 'valgrind:mem', 'valgrind:mem0'].each do |task_name|
     Rake::Task["test:#{task_name}"].prerequisites << :compile

data/build_all

@@ -2,26 +2,32 @@
 #
 #  script to build gems for all relevant platforms:
 #  - MRI et al (standard gem)
-#  - windows (x86-mingw32 and x86-msin32-60)
+#  - windows (x86-mingw32 and x86-mswin32-60)
 #  - jruby
 #
-#  prerequisite is the mingw32 packages.
-#  on ubuntu, `sudo apt-get install mingw32`
-#  for others, read up at https://github.com/luislavena/rake-compiler
+#  here's what I recommend for building all the gems:
 #
-#  on ubuntus 11 and later, you may have issues with building
-#  rake-compiler's rubies against openssl v2. Just comment the lines
-#  out from ossl_ssl.c and you'll be fine.
+#   1. set up a vagrant VM guest running ubuntu lucid 32-bit.
+#   2. install rvm, and install 1.8.7, 1.9.3 and jruby.
+#   3. `sudo apt-get install mingw32`
 #
-#  you may have issues with Pathname conversion to String in bundler. Add this to the offending bundler file:
+#  as you build, you may run into these problems:
 #
-#     class Pathname
-#       def to_str
-#         to_s
-#       end
-#     end
+#  - on ubuntus 11 and later, you may have issues with building
+#    rake-compiler's rubies against openssl v2. Just comment the lines
+#    out from ossl_ssl.c and you'll be fine.
 #
-#  you may also have to remove some of the symbol checks in extconf.rb ... crazy.
+#  - you may have issues with Pathname conversion to String in
+#    bundler. Add this to the offending bundler file:
+#
+#      class Pathname
+#        def to_str
+#          to_s
+#        end
+#      end
+#
+#  - you may also have to hack rubygems.rb to eliminate a reference to
+#    RUBY_ENGINE
 #
 
 HOST=
@@ -53,16 +59,23 @@ rm -rf gems
 mkdir -p gems
 
 # windows
-# rvm_use 1.8.7
-# if [[ ! -a /home/mike/.rake-compiler/ruby/ruby-1.8.7-p330/lib/ruby/1.8.7/x86_64-linux/rbconfig.rb ]] ; then
-#     bundle exec rake-compiler cross-ruby VERSION=1.8.7-p330
-# fi
-# if [[ ! -a /home/mike/.rake-compiler/ruby/ruby-1.9.2-p136/lib/ruby/1.9.1/x86_64-linux/rbconfig.rb ]] ; then
-#     bundle exec rake-compiler cross-ruby VERSION=1.9.2-p136
-# fi
-# bundle exec rake cross
-# bundle exec rake gem:windows
-# cp -v pkg/nokogiri*x86-{mingw32,mswin32}*.gem gems
+platform=$(uname -i)
+if [[ $platform =~ "64" ]] ; then
+    echo ""
+    echo "ERROR: You need to build the windows gem on a 32-bit machine!"
+    echo ""
+    exit 1
+fi
+rvm_use 1.8.7
+if [[ ! -a ${HOME}/.rake-compiler/ruby/ruby-1.8.7-p330/lib/ruby/1.8.7/x86_64-linux/rbconfig.rb ]] ; then
+    bundle exec rake-compiler cross-ruby VERSION=1.8.7-p330
+fi
+if [[ ! -a ${HOME}/.rake-compiler/ruby/ruby-1.9.2-p136/lib/ruby/1.9.1/x86_64-linux/rbconfig.rb ]] ; then
+    bundle exec rake-compiler cross-ruby VERSION=1.9.2-p136
+fi
+bundle exec rake cross
+bundle exec rake gem:windows
+cp -v pkg/nokogiri*x86-{mingw32,mswin32}*.gem gems
 
 # MRI
 rvm_use 1.8.7
@@ -70,11 +83,11 @@ bundle exec rake gem
 cp -v pkg/nokogiri*.gem gems # should only be one at this point in the script
 
 # jruby
-rvm_use jruby-1.6.5
+rvm_use jruby
 bundle install --quiet --local || bundle install
 bundle exec rake clean clobber
 rvm_use 1.8.7
 bundle exec rake generate
-rvm_use jruby-1.6.5
+rvm_use jruby
 bundle exec rake gem
-cp -v pkg/nokogiri*java.gem gems
+cp -v pkg/nokogiri*java.gem gems

data/ext/java/nokogiri/HtmlDocument.java

@@ -55,6 +55,10 @@ import org.w3c.dom.NodeList;
  */
 @JRubyClass(name="Nokogiri::HTML::Document", parent="Nokogiri::XML::Document")
 public class HtmlDocument extends XmlDocument {
+    private static final String DEFAULT_CONTENT_TYPE = "html";
+    private static final String DEFAULT_PUBLIC_ID = "-//W3C//DTD HTML 4.01//EN";
+    private static final String DEFAULT_SYTEM_ID = "http://www.w3.org/TR/html4/strict.dtd";
+
     private String parsed_encoding = null;
 
     public HtmlDocument(Ruby ruby, RubyClass klazz) {
@@ -82,6 +86,28 @@ public class HtmlDocument extends XmlDocument {
         return htmlDocument;
     }
 
+    public IRubyObject getInternalSubset(ThreadContext context) {
+        IRubyObject internalSubset = super.getInternalSubset(context);
+
+        // html documents are expected to have a default internal subset
+        // the default values are the same ones used when the following
+        // feature is turned on
+        // "http://cyberneko.org/html/features/insert-doctype"
+        // the reason we don't turn it on, is because it overrides the document's
+        // declared doctype declaration.
+
+        if (internalSubset.isNil()) {
+            internalSubset = XmlDtd.newEmpty(context.getRuntime(),
+                                             getDocument(),
+                                             context.getRuntime().newString(DEFAULT_CONTENT_TYPE),
+                                             context.getRuntime().newString(DEFAULT_PUBLIC_ID),
+                                             context.getRuntime().newString(DEFAULT_SYTEM_ID));
+            setInternalSubset(internalSubset);
+        }
+
+        return internalSubset;
+    }
+
     public static IRubyObject do_parse(ThreadContext context,
                                        IRubyObject klass,
                                        IRubyObject[] args) {

data/ext/java/nokogiri/XmlDocument.java

@@ -64,6 +64,7 @@ import org.jruby.runtime.ThreadContext;
 import org.jruby.runtime.builtin.IRubyObject;
 import org.w3c.dom.Attr;
 import org.w3c.dom.Document;
+import org.w3c.dom.DocumentType;
 import org.w3c.dom.NamedNodeMap;
 import org.w3c.dom.Node;
 import org.w3c.dom.NodeList;
@@ -288,7 +289,7 @@ public class XmlDocument extends XmlNode {
 
     @JRubyMethod
     public IRubyObject encoding(ThreadContext context) {
-        if (this.encoding == null) {
+        if (this.encoding == null || this.encoding.isNil()) {
             if (getDocument().getXmlEncoding() == null) {
                 this.encoding = context.getRuntime().getNil();
             } else {
@@ -296,7 +297,7 @@ public class XmlDocument extends XmlNode {
             }
         }
 
-        return this.encoding;
+        return this.encoding.isNil() ? this.encoding : this.encoding.asString().encode(context, context.getRuntime().newString("UTF-8"));
     }
 
     @JRubyMethod(meta = true)
@@ -438,8 +439,20 @@ public class XmlDocument extends XmlNode {
         IRubyObject dtd = (IRubyObject) node.getUserData(DTD_INTERNAL_SUBSET);
 
         if (dtd == null) {
-            if (getDocument().getDoctype() == null) dtd = context.getRuntime().getNil();
-            else dtd = XmlDtd.newFromInternalSubset(context.getRuntime(), getDocument());
+            Document document = getDocument();
+            if (document.getUserData(XmlDocument.DTD_RAW_DOCUMENT) != null) {
+                dtd = XmlDtd.newFromInternalSubset(context.getRuntime(), document);
+            } else if (document.getDoctype() != null) {
+                DocumentType docType = document.getDoctype();
+                dtd = XmlDtd.newEmpty(context.getRuntime(),
+                                      document,
+                                      context.getRuntime().newString(docType.getName()),
+                                      context.getRuntime().newString(docType.getPublicId()),
+                                      context.getRuntime().newString(docType.getSystemId()));
+            } else {
+                dtd = context.getRuntime().getNil();
+            }
+
             setInternalSubset(dtd);
         }
 

data/ext/java/nokogiri/XmlDocumentFragment.java

@@ -33,7 +33,6 @@
 package nokogiri;
 
 import static nokogiri.internals.NokogiriHelpers.getLocalNameForNamespace;
-import static nokogiri.internals.NokogiriHelpers.getLocalPart;
 import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
 import static nokogiri.internals.NokogiriHelpers.getPrefix;
 import static nokogiri.internals.NokogiriHelpers.isNamespace;
@@ -93,7 +92,6 @@ public class XmlDocumentFragment extends XmlNode {
         if (args.length > 1 && args[1] instanceof RubyString) {
             args[1] = trim(context, doc, (RubyString)args[1]);
             if (XmlDocumentFragment.isTag((RubyString)args[1])) {
-                args[1] = RubyString.newString(context.getRuntime(), ignoreNamespaceIfNeeded(doc, rubyStringToString(args[1])));
                 args[1] = RubyString.newString(context.getRuntime(), addNamespaceDeclIfNeeded(doc, rubyStringToString(args[1])));
             }
         }
@@ -127,46 +125,10 @@ public class XmlDocumentFragment extends XmlNode {
         if (str.startsWith("<") && str.endsWith(">")) return true;
         return false;
     }
-    
+
     private static Pattern qname_pattern = Pattern.compile("[^</:>\\s]+:[^</:>=\\s]+");
     private static Pattern starttag_pattern = Pattern.compile("<[^</>]+>");
 
-    /**
-     * Remove the namespace from @param tag and/or the attributes in @param tag if
-     * the namespace is the default in the current context as defined by @param doc.
-     *
-     * @return the normalized tag.
-     */
-    private static String ignoreNamespaceIfNeeded(XmlDocument doc, String tags) {
-        if (doc.getDocument() == null) return tags;
-        Map<String, String> rewriteTable = new HashMap<String, String>();
-        // we have to make sure that we don't replace strings in double quotes, e.g.
-        // 'urn:xmpp:foospec:barfoo' in '<foobar xmlns="urn:xmpp:foospec:barfoo"/>'
-        // has to remain the same
-        String[] parts = tags.split("\"");
-        for (int partidx = 0; partidx < parts.length; partidx++) {
-          if (partidx % 2 == 1)
-            continue;
-          Matcher matcher = qname_pattern.matcher(parts[partidx]);
-          while(matcher.find()) {
-            String qName = matcher.group();
-            if (doc.getDocument().getDocumentElement() != null) {
-              NamedNodeMap nodeMap = doc.getDocument().getDocumentElement().getAttributes();
-              if (!isNamespaceDefined(qName, nodeMap)) {
-                rewriteTable.put(qName, getLocalPart(qName));
-              }
-            } else {
-              rewriteTable.put(qName, getLocalPart(qName));
-            }
-          }
-        }
-        Set<String> keys = rewriteTable.keySet();
-        for (String key : keys) {
-            tags = tags.replace(key, rewriteTable.get(key));
-        }
-        return tags;
-    }
-    
     private static boolean isNamespaceDefined(String qName, NamedNodeMap nodeMap) {
         if (isNamespace(qName.intern())) return true;
         for (int i=0; i < nodeMap.getLength(); i++) {

data/ext/java/nokogiri/XmlNode.java

@@ -448,6 +448,7 @@ public class XmlNode extends RubyObject {
     public void relink_namespace(ThreadContext context) {
         if (node instanceof Element) {
             Element e = (Element) node;
+            e.getOwnerDocument().setStrictErrorChecking(false);
             e.getOwnerDocument().renameNode(e, e.lookupNamespaceURI(e.getPrefix()), e.getNodeName());
 
             if (e.hasAttributes()) {
@@ -550,7 +551,7 @@ public class XmlNode extends RubyObject {
             str = NokogiriHelpers.getLocalPart(str);
         }
         if (str == null) str = "";
-        name = context.getRuntime().newString(str);
+        name = NokogiriHelpers.stringOrBlank(context.getRuntime(), str);
         return name;
     }
 
@@ -815,7 +816,7 @@ public class XmlNode extends RubyObject {
             if (node == null) {
                 textContent = "";
             } else {
-                textContent = ((Document)this.node).getDocumentElement().getTextContent().trim();
+                textContent = ((Document)this.node).getDocumentElement().getTextContent();
             }
         } else {
             textContent = this.node.getTextContent();

data/ext/java/nokogiri/XmlSaxPushParser.java

@@ -35,16 +35,16 @@ package nokogiri;
 import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
 import static org.jruby.javasupport.util.RuntimeHelpers.invoke;
 
+import java.io.ByteArrayInputStream;
 import java.io.IOException;
-import java.io.InputStream;
-import java.io.OutputStream;
-import java.nio.channels.Channels;
-import java.nio.channels.Pipe;
 import java.util.concurrent.Callable;
 import java.util.concurrent.ExecutorService;
 import java.util.concurrent.Executors;
+import java.util.concurrent.Future;
 import java.util.concurrent.FutureTask;
 
+import nokogiri.internals.ClosedStreamException;
+import nokogiri.internals.NokogiriBlockingQueueInputStream;
 import nokogiri.internals.ParserContext;
 
 import org.jruby.Ruby;
@@ -69,8 +69,7 @@ public class XmlSaxPushParser extends RubyObject {
     ParserContext.Options options;
     IRubyObject optionsRuby;
     IRubyObject saxParser;
-    OutputStream ostream = null;
-    InputStream istream = null;
+    NokogiriBlockingQueueInputStream stream;
     ParserTask parserTask = null;
     FutureTask<XmlSaxParserContext> futureTask = null;
     ExecutorService executor = null;
@@ -120,53 +119,43 @@ public class XmlSaxPushParser extends RubyObject {
         byte[] data = null;
         if (chunk instanceof RubyString || chunk.respondsTo("to_str")) {
             data = chunk.convertToString().getBytes();
-        } else { 
-            try {
-                terminateTask();
-            } catch (IOException e) {
-                throw context.getRuntime().newRuntimeError(e.getMessage());
-            }
-            XmlSyntaxError xmlSyntaxError = 
+        } else {
+            terminateTask(context);
+            XmlSyntaxError xmlSyntaxError =
                 (XmlSyntaxError) NokogiriService.XML_SYNTAXERROR_ALLOCATOR.allocate(context.getRuntime(), getNokogiriClass(context.getRuntime(), "Nokogiri::XML::SyntaxError"));
             throw new RaiseException(xmlSyntaxError);
         }
 
         int errorCount0 = parserTask.getErrorCount();;
-        
-        try {
-            if (isLast.isTrue()) {
-                IRubyObject document = invoke(context, this, "document");
-                invoke(context, document, "end_document");
-                terminateTask();
-            } else {
-                ostream.write(data);
-                Thread.currentThread().sleep(10);   // gives a reader a chance to work
+
+
+        if (isLast.isTrue()) {
+            IRubyObject document = invoke(context, this, "document");
+            invoke(context, document, "end_document");
+            terminateTask(context);
+        } else {
+            try {
+              Future<Void> task = stream.addChunk(new ByteArrayInputStream(data));
+              task.get();
+            } catch (ClosedStreamException ex) {
+              // this means the stream is closed, ignore this exception
+            } catch (Exception e) {
+              throw context.getRuntime().newRuntimeError(e.getMessage());
             }
-        } catch (IOException e) {
-            throw context.getRuntime().newRuntimeError(e.getMessage());
-        } catch (InterruptedException e) {
-            throw context.getRuntime().newRuntimeError(e.getMessage());
+
         }
 
         if (!options.recover && parserTask.getErrorCount() > errorCount0) {
-            try {
-                terminateTask();
-            } catch (IOException e) {
-                throw context.getRuntime().newRuntimeError(e.getMessage());
-            }
+            terminateTask(context);
             throw new RaiseException(parserTask.getLastError(), true);
         }
 
         return this;
     }
-    
+
     private void initialize_task(ThreadContext context) throws IOException {
-        if (futureTask == null || ostream == null || istream == null) {
-            Pipe pipe = Pipe.open();
-            Pipe.SinkChannel sink = pipe.sink();
-            ostream = Channels.newOutputStream(sink);
-            Pipe.SourceChannel source = pipe.source();
-            istream = Channels.newInputStream(source);
+        if (futureTask == null || stream == null) {
+            stream = new NokogiriBlockingQueueInputStream();
 
             parserTask = new ParserTask(context, saxParser);
             futureTask = new FutureTask<XmlSaxParserContext>(parserTask);
@@ -174,40 +163,53 @@ public class XmlSaxPushParser extends RubyObject {
             executor.submit(futureTask);
         }
     }
-    
-    private synchronized void terminateTask() throws IOException {
+
+    private synchronized void terminateTask(ThreadContext context) {
+        try {
+          Future<Void> task = stream.addChunk(NokogiriBlockingQueueInputStream.END);
+          task.get();
+        } catch (ClosedStreamException ex) {
+          // ignore this exception, it means the stream was closed
+        } catch (Exception e) {
+            throw context.getRuntime().newRuntimeError(e.getMessage());
+        }
         futureTask.cancel(true);
         executor.shutdown();
-        ostream.close();
-        istream.close();
-        ostream = null;
-        istream = null;
+        executor = null;
+        stream = null;
+        futureTask = null;
     }
-    
+
     private class ParserTask implements Callable<XmlSaxParserContext> {
-        private ThreadContext context;
-        private IRubyObject handler;
-        private XmlSaxParserContext parser;
-        
+        private final ThreadContext context;
+        private final IRubyObject handler;
+        private final XmlSaxParserContext parser;
+
         private ParserTask(ThreadContext context, IRubyObject handler) {
             RubyClass klazz = getNokogiriClass(context.getRuntime(), "Nokogiri::XML::SAX::ParserContext");
             this.context = context;
             this.handler = handler;
-            this.parser = (XmlSaxParserContext) XmlSaxParserContext.parse_stream(context, klazz, istream);
+            this.parser = (XmlSaxParserContext) XmlSaxParserContext.parse_stream(context, klazz, stream);
         }
 
         @Override
         public XmlSaxParserContext call() throws Exception {
+          try {
             parser.parse_with(context, handler);
-            return parser;
+          } finally {
+            // we have to close the stream before exiting, otherwise someone
+            // can add a chunk and block on task.get() forever.
+            stream.close();
+          }
+          return parser;
         }
-        
+
         private synchronized int getErrorCount() {
             // check for null because thread may not have started yet
             if (parser.getNokogiriHandler() == null) return 0;
             else return parser.getNokogiriHandler().getErrorCount();
         }
-        
+
         private synchronized RubyException getLastError() {
             return (RubyException) parser.getNokogiriHandler().getLastError();
         }