nokogiri 1.6.6.4-java → 1.6.7-java

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

@@ -37,8 +37,8 @@ import static nokogiri.internals.NokogiriHelpers.getPrefix;
 import static nokogiri.internals.NokogiriHelpers.isNamespace;
 import static nokogiri.internals.NokogiriHelpers.stringOrNil;
 
-import java.util.ArrayDeque;
 import java.util.LinkedList;
+import java.util.Stack;
 
 import nokogiri.XmlSyntaxError;
 
@@ -62,7 +62,7 @@ import org.xml.sax.ext.DefaultHandler2;
  * @author Yoko Harada <yokolet@gmail.com>
  */
 public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
-    private StringBuffer buffer;
+    Stack<StringBuffer> characterStack;
     private final Ruby ruby;
     private final RubyClass attrClass;
     private final IRubyObject object;
@@ -100,6 +100,7 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
     @Override
     public void startDocument() throws SAXException {
         call("start_document");
+        characterStack = new Stack();
     }
 
     @Override
@@ -111,6 +112,13 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
 
     @Override
     public void endDocument() throws SAXException {
+        StringBuffer sb;
+        if (!characterStack.empty()) {
+            for (int i=0; i<characterStack.size(); i++) {
+                sb = characterStack.get(i);
+                call("characters", ruby.newString(sb.toString()));
+            }
+        }
         call("end_document");
     }
 
@@ -120,12 +128,10 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
     }
 
     /*
-     * This has to call either "start_element" or
-     * "start_element_namespace" depending on whether there are any
-     * namespace attributes.
+     * This calls "start_element_namespace".
      *
      * Attributes that define namespaces are passed in a separate
-     * array of of <code>[:prefix, :uri]</code> arrays and are not
+     * array of <code>[:prefix, :uri]</code> arrays and are not
      * passed with the other attributes.
      */
     @Override
@@ -150,7 +156,7 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
 
             if (isNamespace(qn) && !fromFragmentHandler) {
                 // I haven't figured the reason out yet, but, in somewhere,
-                // namespace is converted to array in array in array and cause
+                // namespace is converted to array in array and cause
                 // TypeError at line 45 in fragment_handler.rb
                 RubyArray ns = RubyArray.newArray(ruby, 2);
                 if (ln.equals("xmlns")) ln = null;
@@ -187,6 +193,7 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
              stringOrNil(ruby, getPrefix(qName)),
              stringOrNil(ruby, uri),
              rubyNSAttr);
+        characterStack.push(new StringBuffer());
     }
     
     private static String[] emptyAttrs =
@@ -224,6 +231,8 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
 
     @Override
     public void endElement(String uri, String localName, String qName) throws SAXException {
+        StringBuffer sb = characterStack.pop();
+        call("characters", ruby.newString(sb.toString()));
         call("end_element_namespace",
              stringOrNil(ruby, localName),
              stringOrNil(ruby, getPrefix(qName)),
@@ -232,11 +241,8 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
 
     @Override
     public void characters(char[] ch, int start, int length) throws SAXException {
-        if (buffer != null) {
-          buffer.append(new String(ch, start, length));
-        } else {
-          call("characters", ruby.newString(new String(ch, start, length)));
-        }
+        StringBuffer sb = characterStack.peek();
+        sb.append(new String(ch, start, length));
     }
 
     @Override
@@ -246,13 +252,13 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
 
     @Override
     public void startCDATA() throws SAXException {
-        buffer = new StringBuffer();
+        characterStack.push(new StringBuffer());
     }
 
     @Override
     public void endCDATA() throws SAXException {
-        call("cdata_block", ruby.newString(buffer.toString()));
-        buffer = null;
+        StringBuffer sb = characterStack.pop();
+        call("cdata_block", ruby.newString(sb.toString()));
     }
 
     @Override

data/ext/java/nokogiri/internals/ParserContext.java

@@ -100,25 +100,29 @@ public class ParserContext extends RubyObject {
         RubyString stringData = null;
         if (invoke(context, data, "respond_to?",
                    ruby.newSymbol("to_io").to_sym()).isTrue()) {
-            /* IO or other object that responds to :to_io */
             RubyIO io =
                 (RubyIO) TypeConverter.convertToType(data,
                                                      ruby.getIO(),
                                                      "to_io");
             // use unclosedable input stream to fix #495
             source.setByteStream(new UncloseableInputStream(io.getInStream()));
-        } else {
-            if (invoke(context, data, "respond_to?",
+
+        } else if (invoke(context, data, "respond_to?",
+                          ruby.newSymbol("read").to_sym()).isTrue()) {
+            stringData = invoke(context, data, "read").convertToString();
+
+        } else if (invoke(context, data, "respond_to?",
                           ruby.newSymbol("string").to_sym()).isTrue()) {
-                /* StringIO or other object that responds to :string */
-                stringData = invoke(context, data, "string").convertToString();
-            } else if (data instanceof RubyString) {
-                stringData = (RubyString) data;
-            } else {
-                throw ruby.newArgumentError(
-                    "must be kind_of String or respond to :to_io or :string");
-            }
+            stringData = invoke(context, data, "string").convertToString();
+
+        } else if (data instanceof RubyString) {
+            stringData = (RubyString) data;
+
+        } else {
+            throw ruby.newArgumentError(
+                "must be kind_of String or respond to :to_io, :read, or :string");
         }
+
         if (stringData != null) {
             String encName = null;
             if (stringData.encoding(context) != null) {

data/ext/nokogiri/extconf.rb

@@ -148,8 +148,8 @@ def iconv_configure_flags
 
       return [
         '--with-iconv=yes',
-        *("CPPFLAGS=#{idirs.map { |dir| '-I' << dir }.join(' ')}".quote if idirs),
-        *("LDFLAGS=#{ldirs.map { |dir| '-L' << dir }.join(' ')}".quote if ldirs),
+        *("CPPFLAGS=#{idirs.map { |dir| '-I' << dir }.join(' ')}" if idirs),
+        *("LDFLAGS=#{ldirs.map { |dir| '-L' << dir }.join(' ')}" if ldirs),
       ]
     end
   }
@@ -164,22 +164,32 @@ def iconv_configure_flags
 
     return [
       '--with-iconv=yes',
-      "CPPFLAGS=#{cflags}".quote,
-      "LDFLAGS=#{ldflags}".quote,
-      "LIBS=#{libs}".quote,
+      "CPPFLAGS=#{cflags}",
+      "LDFLAGS=#{ldflags}",
+      "LIBS=#{libs}",
     ]
   end
 
   asplode "libiconv"
 end
 
+# When using rake-compiler-dock on Windows, the underlying Virtualbox shared
+# folders don't support symlinks, but libiconv expects it for a build on
+# Linux. We work around this limitation by using the temp dir for cooking.
+def chdir_for_build
+  build_dir = ENV['RCD_HOST_RUBY_PLATFORM'].to_s =~ /mingw|mswin|cygwin/ ? '/tmp' : '.'
+  Dir.chdir(build_dir) do
+    yield
+  end
+end
+
 def process_recipe(name, version, static_p, cross_p)
   MiniPortile.new(name, version).tap do |recipe|
     recipe.target = portsdir = File.join(ROOT, "ports")
     # Prefer host_alias over host in order to use i586-mingw32msvc as
     # correct compiler prefix for cross build, but use host if not set.
     recipe.host = RbConfig::CONFIG["host_alias"].empty? ? RbConfig::CONFIG["host"] : RbConfig::CONFIG["host_alias"]
-    recipe.patch_files = Dir[File.join(portsdir, "patches", name, "*.patch")].sort
+    recipe.patch_files = Dir[File.join(ROOT, "patches", name, "*.patch")].sort
 
     yield recipe
 
@@ -190,7 +200,7 @@ def process_recipe(name, version, static_p, cross_p)
     recipe.configure_options.flatten!
 
     recipe.configure_options.delete_if { |option|
-      case option.shellsplit.first
+      case option
       when /\A(\w+)=(.*)\z/
         env[$1] = $2
         true
@@ -221,14 +231,14 @@ def process_recipe(name, version, static_p, cross_p)
 
     if RbConfig::CONFIG['target_cpu'] == 'universal'
       %w[CFLAGS LDFLAGS].each { |key|
-        unless env[key].shellsplit.include?('-arch')
+        unless env[key].include?('-arch')
           env[key] << ' ' << RbConfig::CONFIG['ARCH_FLAG']
         end
       }
     end
 
     recipe.configure_options += env.map { |key, value|
-      "#{key}=#{value}".shellescape
+      "#{key}=#{value}"
     }
 
     message <<-"EOS"
@@ -275,7 +285,9 @@ versions of libxml2 provided by OS/package vendors.
 
     checkpoint = "#{recipe.target}/#{recipe.name}-#{recipe.version}-#{recipe.host}.installed"
     unless File.exist?(checkpoint)
-      recipe.cook
+      chdir_for_build do
+        recipe.cook
+      end
       FileUtils.touch checkpoint
     end
     recipe.activate
@@ -301,19 +313,6 @@ end
 # https://bugs.ruby-lang.org/issues/8074
 @libdir_basename = "lib" if RUBY_VERSION < '2.1.0'
 
-# Workaround for #1102
-def monkey_patch_mini_portile
-  MiniPortile.class_eval do
-    def patch
-      @patch_files.each do |full_path|
-        next unless File.exists?(full_path)
-        output "Running patch with #{full_path}..."
-        execute('patch', %Q(patch -p1 < "#{full_path}"))
-      end
-    end
-  end
-end
-
 #
 # main
 #
@@ -392,8 +391,12 @@ when using_system_libraries?
 else
   message "Building nokogiri using packaged libraries.\n"
 
-  require 'mini_portile'
-  monkey_patch_mini_portile
+  # The gem version constraint in the Rakefile is not respected at install time.
+  # Keep this version in sync with the one in the Rakefile !
+  gem "mini_portile2", "~> 2.0.0.rc2"
+  require 'mini_portile2'
+  message "Using mini_portile version #{MiniPortile::VERSION}\n"
+
   require 'yaml'
 
   static_p = enable_config('static', true) or
@@ -443,17 +446,17 @@ else
     libiconv_recipe = process_recipe("libiconv", dependencies["libiconv"], static_p, cross_build_p) do |recipe|
       recipe.files = ["http://ftp.gnu.org/pub/gnu/libiconv/#{recipe.name}-#{recipe.version}.tar.gz"]
       recipe.configure_options += [
-        "CPPFLAGS='-Wall'",
-        "CFLAGS='-O2 -g'",
-        "CXXFLAGS='-O2 -g'",
+        "CPPFLAGS=-Wall",
+        "CFLAGS=-O2 -g",
+        "CXXFLAGS=-O2 -g",
         "LDFLAGS="
       ]
     end
   else
-    if darwin_p && !File.exist?('/usr/include/iconv.h')
+    if darwin_p && !have_header('iconv.h')
       abort <<'EOM'.chomp
 -----
-The file "/usr/include/iconv.h" is missing in your build environment,
+The file "iconv.h" is missing in your build environment,
 which means you haven't installed Xcode Command Line Tools properly.
 
 To install Command Line Tools, try running `xcode-select --install` on
@@ -473,7 +476,7 @@ EOM
   end
 
   libxml2_recipe = process_recipe("libxml2", dependencies["libxml2"], static_p, cross_build_p) do |recipe|
-    recipe.files = ["ftp://ftp.xmlsoft.org/libxml2/#{recipe.name}-#{recipe.version}.tar.gz"]
+    recipe.files = ["http://xmlsoft.org/sources/#{recipe.name}-#{recipe.version}.tar.gz"]
     recipe.configure_options += [
       "--without-python",
       "--without-readline",
@@ -485,7 +488,7 @@ EOM
   end
 
   libxslt_recipe = process_recipe("libxslt", dependencies["libxslt"], static_p, cross_build_p) do |recipe|
-    recipe.files = ["ftp://ftp.xmlsoft.org/libxml2/#{recipe.name}-#{recipe.version}.tar.gz"]
+    recipe.files = ["http://xmlsoft.org/sources/#{recipe.name}-#{recipe.version}.tar.gz"]
     recipe.configure_options += [
       "--without-python",
       "--without-crypto",
@@ -526,8 +529,8 @@ EOM
       }
 
       # Defining a macro that expands to a C string; double quotes are significant.
-      $CPPFLAGS << ' ' << "-DNOKOGIRI_#{recipe.name.upcase}_PATH=\"#{recipe.path}\"".shellescape
-      $CPPFLAGS << ' ' << "-DNOKOGIRI_#{recipe.name.upcase}_PATCHES=\"#{recipe.patch_files.map { |path| File.basename(path) }.join(' ')}\"".shellescape
+      $CPPFLAGS << ' ' << "-DNOKOGIRI_#{recipe.name.upcase}_PATH=\"#{recipe.path}\"".inspect
+      $CPPFLAGS << ' ' << "-DNOKOGIRI_#{recipe.name.upcase}_PATCHES=\"#{recipe.patch_files.map { |path| File.basename(path) }.join(' ')}\"".inspect
 
       case libname
       when 'xml2'

data/ext/nokogiri/xml_node.c

@@ -778,11 +778,19 @@ static VALUE namespaced_key_eh(VALUE self, VALUE attribute, VALUE namespace)
  *
  * Set the +property+ to +value+
  */
-static VALUE set(VALUE self, VALUE property, VALUE value)
+static VALUE set(VALUE node_rb, VALUE property_name_rb, VALUE property_value_rb)
 {
   xmlNodePtr node, cur;
-  xmlAttrPtr prop;
-  Data_Get_Struct(self, xmlNode, node);
+  xmlChar*   property_name ;
+  xmlAttrPtr property;
+
+  Data_Get_Struct(node_rb, xmlNode, node);
+
+  if (node->type != XML_ELEMENT_NODE) {
+    return(Qnil); // TODO: would raising an exception be more appropriate?
+  }
+
+  property_name = (xmlChar *)StringValuePtr(property_name_rb);
 
   /* If a matching attribute node already exists, then xmlSetProp will destroy
    * the existing node's children. However, if Nokogiri has a node object
@@ -790,11 +798,9 @@ static VALUE set(VALUE self, VALUE property, VALUE value)
    *
    * We can avoid this by unlinking these nodes first.
    */
-  if (node->type != XML_ELEMENT_NODE)
-    return(Qnil);
-  prop = xmlHasProp(node, (xmlChar *)StringValuePtr(property));
-  if (prop && prop->children) {
-    for (cur = prop->children; cur; cur = cur->next) {
+  property = xmlHasProp(node, property_name);
+  if (property && property->children) {
+    for (cur = property->children; cur; cur = cur->next) {
       if (cur->_private) {
         nokogiri_root_node(cur);
         xmlUnlinkNode(cur);
@@ -802,10 +808,14 @@ static VALUE set(VALUE self, VALUE property, VALUE value)
     }
   }
 
-  xmlSetProp(node, (xmlChar *)StringValuePtr(property),
-      (xmlChar *)StringValuePtr(value));
+  xmlResetLastError();
+  xmlSetStructuredErrorFunc(NULL, Nokogiri_error_silencer);
+
+  xmlSetProp(node, property_name, (xmlChar *)StringValuePtr(property_value_rb));
+
+  xmlSetStructuredErrorFunc(NULL, NULL);
 
-  return value;
+  return property_value_rb;
 }
 
 /*

data/ext/nokogiri/xml_sax_parser_context.c

@@ -206,7 +206,7 @@ static VALUE column(VALUE self)
  *  recovery=(boolean)
  *
  * Should this parser recover from structural errors? It will not stop processing
- * file on structural errors if if set to true
+ * file on structural errors if set to true
  */
 static VALUE set_recovery(VALUE self, VALUE value)
 {
@@ -226,7 +226,7 @@ static VALUE set_recovery(VALUE self, VALUE value)
  *  recovery
  *
  * Should this parser recover from structural errors? It will not stop processing
- * file on structural errors if if set to true
+ * file on structural errors if set to true
  */
 static VALUE get_recovery(VALUE self)
 {

data/lib/nokogiri/version.rb

@@ -1,6 +1,6 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.6.6.4'
+  VERSION = '1.6.7'
 
   class VersionInfo # :nodoc:
     def jruby?

data/lib/nokogiri/xml/node.rb

@@ -32,6 +32,22 @@ module Nokogiri
     # * Nokogiri::XML::Node#next
     # * Nokogiri::XML::Node#previous
     #
+    #
+    # When printing or otherwise emitting a document or a node (and
+    # its subtree), there are a few methods you might want to use:
+    #
+    # * content, text, inner_text, to_str: emit plaintext
+    #  
+    #   These methods will all emit the plaintext version of your
+    #   document, meaning that entities will be replaced (e.g., "<"
+    #   will be replaced with "<"), meaning that any sanitizing will
+    #   likely be un-done in the output.
+    #
+    # * to_s, to_xml, to_html, inner_html: emit well-formed markup
+    #
+    #   These methods will all emit properly-escaped markup, meaning
+    #   that it's suitable for consumption by browsers, parsers, etc.
+    #
     # You may search this node's subtree using Searchable#xpath and Searchable#css
     class Node
       include Nokogiri::XML::PP::Node

data/test/html/test_document.rb

@@ -14,6 +14,11 @@ module Nokogiri
         assert @html.xpath(nil)
       end
 
+      def test_does_not_fail_with_illformatted_html
+        doc = Nokogiri::HTML('"</html>";'.force_encoding(Encoding::BINARY))
+        assert_not_nil doc
+      end
+
       def test_exceptions_remove_newlines
         errors = @html.errors
         assert errors.length > 0, 'has errors'
@@ -88,9 +93,9 @@ module Nokogiri
       def test_document_parse_method_with_url
         require 'open-uri'
         begin
-          html = open('http://google.com').read
-        rescue
-          skip("This test needs the internet. Skips if no internet available.")
+          html = open('https://www.yahoo.com').read
+        rescue Exception => e
+          skip("This test needs the internet. Skips if no internet available. (#{e})")
         end
         doc = Nokogiri::HTML html ,"http:/foobar.foobar/"
         refute_empty doc.to_s, "Document should not be empty"
@@ -637,6 +642,24 @@ eohtml
         assert_equal original_errors.length+1, doc1.errors.length
         assert_match(/ID unique already defined/, doc1.errors.last.to_s)
       end
+
+      def test_silencing_nonparse_errors_during_attribute_insertion_1262
+        # see https://github.com/sparklemotion/nokogiri/issues/1262
+        #
+        # libxml2 emits a warning when this happens; the JRuby
+        # implementation does not. so rather than capture the error in
+        # doc.errors in a platform-dependent way, I'm opting to have
+        # the error silenced.
+        #
+        # So this test doesn't look meaningful, but we want to avoid
+        # having `ID unique-issue-1262 already defined` emitted to
+        # stderr when running the test suite.
+        #
+        doc = Nokogiri::HTML::Document.new
+        Nokogiri::XML::Element.new("div", doc).set_attribute('id', 'unique-issue-1262')
+        Nokogiri::XML::Element.new("div", doc).set_attribute('id', 'unique-issue-1262')
+        assert_equal 0, doc.errors.length
+      end
     end
   end
 end