nokogiri 1.1.0 → 1.1.1

data/History.ja.txt

@@ -1,16 +1,34 @@
-=== HEAD
+=== 1.1.1
 
 * 新しい機能
+  * XML::Node#elem? を追加
+  * XML::Node#attribute_nodes を追加
+  * XML::Attr を追加
+  * XML::Node#delete を追加
+  * XML::NodeSet#inner_html を追加
 
-  * カスタム XPath 機能はある。（ 参照 Nokogiri::XML::Node#xpath ）
-  * カスタム CSS 擬似クラスと機能はある。（ 参照 Nokogiri::XML::Node#css ）
-  * Nokogiri::XML::Node#<< は参照に子ノードを添える
+* バグの修正
+
+  * HTML のノードを \r のエンティティを含まない
+  * CSS::SelectorHandler と XML::XPathHandler を除去
+  * XML::Node#attributes が Attr node を value値に返す
+  * XML::NodeSet が to_xml へ実行
+
+=== 1.1.0
+
+* 新しい機能
+
+  * カスタム XPath 機能はある。（ Nokogiri::XML::Node#xpath 参照 ）
+  * カスタム CSS 擬似クラスと機能はある。（ Nokogiri::XML::Node#css 参照 ）
+  * Nokogiri::XML::Node#<< が作成中に子ノードを自動追加
 
 * バグの修正
 
   * mutex が CSS のキャッシュのアクセスをロックする
-  * GCC 3.3.5 のビルドが修正された
-  * XML::Node#to_xml は
+  * GCC 3.3.5 のビルドに関する問題を修正
+  * XML::Node#to_xml が引数indentationを取る 
+  * XML::Node#dup が引数任意のdepthを取る 
+  * XML::Node#add_previous_sibling が新しい兄弟ノードで返す
 
 === 1.0.7
 

data/History.txt

@@ -1,3 +1,20 @@
+=== 1.1.1
+
+* New features
+
+  * Added XML::Node#elem?
+  * Added XML::Node#attribute_nodes
+  * Added XML::Attr
+  * XML::Node#delete added.
+  * XML::NodeSet#inner_html added.
+
+* Bugfixes
+
+  * Not including an HTML entity for \r for HTML nodes.
+  * Removed CSS::SelectorHandler and XML::XPathHandler
+  * XML::Node#attributes returns an Attr node for the value.
+  * XML::NodeSet implements to_xml
+
 === 1.1.0
 
 * New Features

data/Manifest.txt

@@ -47,7 +47,6 @@ lib/nokogiri/css/generated_tokenizer.rb
 lib/nokogiri/css/node.rb
 lib/nokogiri/css/parser.rb
 lib/nokogiri/css/parser.y
-lib/nokogiri/css/selector_handler.rb
 lib/nokogiri/css/syntax_error.rb
 lib/nokogiri/css/tokenizer.rb
 lib/nokogiri/css/tokenizer.rex
@@ -66,6 +65,7 @@ lib/nokogiri/html/sax/parser.rb
 lib/nokogiri/version.rb
 lib/nokogiri/xml.rb
 lib/nokogiri/xml/after_handler.rb
+lib/nokogiri/xml/attr.rb
 lib/nokogiri/xml/before_handler.rb
 lib/nokogiri/xml/builder.rb
 lib/nokogiri/xml/cdata.rb
@@ -86,7 +86,6 @@ lib/nokogiri/xml/text.rb
 lib/nokogiri/xml/xpath.rb
 lib/nokogiri/xml/xpath/syntax_error.rb
 lib/nokogiri/xml/xpath_context.rb
-lib/nokogiri/xml/xpath_handler.rb
 lib/nokogiri/xslt.rb
 lib/nokogiri/xslt/stylesheet.rb
 test/css/test_nthiness.rb
@@ -120,6 +119,7 @@ test/hpricot/test_xml.rb
 test/html/sax/test_parser.rb
 test/html/test_builder.rb
 test/html/test_document.rb
+test/html/test_node.rb
 test/test_convert_xpath.rb
 test/test_css_cache.rb
 test/test_gc.rb
@@ -128,6 +128,7 @@ test/test_nokogiri.rb
 test/test_reader.rb
 test/test_xslt_transforms.rb
 test/xml/sax/test_parser.rb
+test/xml/test_attr.rb
 test/xml/test_builder.rb
 test/xml/test_cdata.rb
 test/xml/test_comment.rb

data/Rakefile

@@ -62,6 +62,8 @@ namespace :gem do
     task :spec => ['build'] do
       File.open("#{HOE.name}.gemspec", 'w') do |f|
         HOE.spec.platform = 'jruby'
+        HOE.spec.files << GENERATED_PARSER
+        HOE.spec.files << GENERATED_TOKENIZER
         HOE.spec.extensions = []
         f.write(HOE.spec.to_ruby)
       end
@@ -91,7 +93,7 @@ file GENERATED_PARSER => "lib/nokogiri/css/parser.y" do |t|
     racc = "#{::Config::CONFIG['bindir']}/racc" if racc.empty?
     sh "#{racc} -o #{t.name} #{t.prerequisites.first}"
   rescue
-    abort "need racc, get the tarball from http://i.loveruby.net/archive/racc/racc-1.4.5-all.tar.gz" 
+    abort "need racc, sudo gem install racc"
   end
 end
 
@@ -242,18 +244,35 @@ def test_suite_cmdline
   cmdline = "ruby -w -I.:lib:ext:test -rtest/unit -e '%w[#{files.join(' ')}].each {|f| require f}'"
 end
 
+class ValgrindTestTask < Rake::TestTask
+  def initialize *args
+    super
+    %w[ ext lib bin test ].each do |dir|
+      self.libs << dir
+    end
+    self.test_files = FileList['test/**/test_*.rb'] +
+      FileList['test/**/*_test.rb']
+    self.verbose = true
+    self.warning = true
+  end
+end
+
+VALGRIND_BASIC_OPTS = "--num-callers=50 --error-limit=no --partial-loads-ok=yes --undef-value-errors=no"
+
+desc "run test suite under valgrind with basic ruby options"
+ValgrindTestTask.new('test:valgrind').extend(Module.new {
+  def ruby *args
+    cmd = "valgrind #{VALGRIND_BASIC_OPTS} #{RUBY} #{args.join(' ')}"
+    puts cmd
+    system cmd
+  end
+})
+Rake::Task['test:valgrind'].prerequisites << :build
+
 namespace :test do
   # partial-loads-ok and undef-value-errors necessary to ignore
   # spurious (and eminently ignorable) warnings from the ruby
   # interpreter
-  VALGRIND_BASIC_OPTS = "--num-callers=50 --error-limit=no --partial-loads-ok=yes --undef-value-errors=no"
-
-  desc "run test suite under valgrind with basic ruby options"
-  task :valgrind => :build do
-    cmdline = "valgrind #{VALGRIND_BASIC_OPTS} #{test_suite_cmdline}"
-    puts cmdline
-    system cmdline
-  end
 
   desc "run test suite under valgrind with memory-fill ruby options"
   task :valgrind_mem => :build do

data/ext/nokogiri/extconf.rb

@@ -19,35 +19,6 @@ $LIBPATH << "/opt/local/lib" if use_macports
 
 $CFLAGS << " -O3 -Wall -Wcast-qual -Wwrite-strings -Wconversion -Wmissing-noreturn -Winline"
 
-if Config::CONFIG['target_os'] == 'mingw32'
-  find_library('xml2', 'xmlParseDoc',
-               File.join(ROOT, 'cross', 'libxml2-2.7.2.win32', 'bin'))
-  find_library('xslt', 'xsltParseStylesheetDoc',
-               File.join(ROOT, 'cross', 'libxslt-1.1.24.win32', 'bin'))
-  find_library('exslt', 'exsltFuncRegister',
-               File.join(ROOT, 'cross', 'libxslt-1.1.24.win32', 'bin'))
-else
-  find_library('xml2', 'xmlParseDoc',
-               LIBDIR,
-               '/opt/local/lib',
-               '/usr/local/lib',
-               '/usr/lib'
-    )
-  find_library('xslt', 'xsltParseStylesheetDoc',
-               LIBDIR,
-               '/opt/local/lib',
-               '/usr/local/lib',
-               '/usr/lib'
-    )
-  find_library('exslt', 'exsltFuncRegister',
-               LIBDIR,
-               '/opt/local/lib',
-               '/usr/local/lib',
-               '/usr/lib'
-    )
-end
-
-
 if Config::CONFIG['target_os'] == 'mingw32'
   header = File.join(ROOT, 'cross', 'libxml2-2.7.2.win32', 'include')
   unless find_header('libxml/xmlversion.h', header)
@@ -89,8 +60,34 @@ else
   unless find_header('libexslt/exslt.h', *HEADER_DIRS)
     abort "need libxslt"
   end
+end
 
-  version = try_constant('LIBXML_VERSION', 'libxml/xmlversion.h')
+if Config::CONFIG['target_os'] == 'mingw32'
+  find_library('xml2', 'xmlParseDoc',
+               File.join(ROOT, 'cross', 'libxml2-2.7.2.win32', 'bin'))
+  find_library('xslt', 'xsltParseStylesheetDoc',
+               File.join(ROOT, 'cross', 'libxslt-1.1.24.win32', 'bin'))
+  find_library('exslt', 'exsltFuncRegister',
+               File.join(ROOT, 'cross', 'libxslt-1.1.24.win32', 'bin'))
+else
+  find_library('xml2', 'xmlParseDoc',
+               LIBDIR,
+               '/opt/local/lib',
+               '/usr/local/lib',
+               '/usr/lib'
+    )
+  find_library('xslt', 'xsltParseStylesheetDoc',
+               LIBDIR,
+               '/opt/local/lib',
+               '/usr/local/lib',
+               '/usr/lib'
+    )
+  find_library('exslt', 'exsltFuncRegister',
+               LIBDIR,
+               '/opt/local/lib',
+               '/usr/local/lib',
+               '/usr/lib'
+    )
 end
 
 create_makefile('nokogiri/native')

data/ext/nokogiri/xml_node.c

@@ -60,10 +60,11 @@ static VALUE internal_subset(VALUE self)
   if(!node->doc) return Qnil;
 
   doc = node->doc;
+  xmlDtdPtr dtd = xmlGetIntSubset(doc);
 
-  if(!doc->intSubset) return Qnil;
+  if(!dtd) return Qnil;
 
-  return Nokogiri_wrap_xml_node((xmlNodePtr)doc->intSubset);
+  return Nokogiri_wrap_xml_node((xmlNodePtr)dtd);
 }
 
 /*
@@ -216,22 +217,6 @@ static VALUE set(VALUE self, VALUE property, VALUE value)
   return value;
 }
 
-/*
- *  call-seq:
- *    remove_attribute(property)
- *
- *  remove the property +property+
- */
-static VALUE remove_prop(VALUE self, VALUE property)
-{
-  xmlNodePtr node;
-  xmlAttrPtr attr ;
-  Data_Get_Struct(self, xmlNode, node);
-  attr = xmlHasProp(node, (xmlChar *)StringValuePtr(property));
-  if (attr) { xmlRemoveProp(attr); }
-  return Qnil;
-}
-
 /*
  * call-seq:
  *   get(attribute)
@@ -252,17 +237,17 @@ static VALUE get(VALUE self, VALUE attribute)
 
 /*
  *  call-seq:
- *    attributes()
+ *    attribute_nodes()
  *
- *  returns a hash containing the node's attributes.
+ *  returns a list containing the Node attributes.
  */
-static VALUE attributes(VALUE self)
+static VALUE attribute_nodes(VALUE self)
 {
     /* this code in the mode of xmlHasProp() */
     xmlNodePtr node ;
     VALUE attr ;
 
-    attr = rb_hash_new() ;
+    attr = rb_ary_new() ;
     Data_Get_Struct(self, xmlNode, node);
 
     Nokogiri_xml_node_properties(node, attr);
@@ -270,6 +255,22 @@ static VALUE attributes(VALUE self)
     return attr ;
 }
 
+
+/*
+ *  call-seq:
+ *    namespace()
+ *
+ *  returns the namespace prefix for the node, if one exists.
+ */
+static VALUE namespace(VALUE self)
+{
+  xmlNodePtr node ;
+  Data_Get_Struct(self, xmlNode, node);
+  if (node->ns && node->ns->prefix)
+    return rb_str_new2((const char *)node->ns->prefix) ;
+  return Qnil ;
+}
+
 /*
  *  call-seq:
  *    namespaces()
@@ -349,9 +350,16 @@ static VALUE add_child(VALUE self, VALUE child)
   Data_Get_Struct(child, xmlNode, node);
   Data_Get_Struct(self, xmlNode, parent);
 
+  xmlUnlinkNode(node) ;
+
   if(!(new_child = xmlAddChild(parent, node)))
     rb_raise(rb_eRuntimeError, "Could not add new child");
 
+  // the child was a text node that was coalesced. we need to have the object
+  // point at SOMETHING, or we'll totally bomb out.
+  if (new_child != node)
+    DATA_PTR(child) = new_child ;
+
   return Nokogiri_wrap_xml_node(new_child);
 }
 
@@ -427,10 +435,16 @@ static VALUE path(VALUE self)
  */
 static VALUE add_next_sibling(VALUE self, VALUE rb_node)
 {
-  xmlNodePtr node, new_sibling;
+  xmlNodePtr node, _new_sibling, new_sibling;
   Data_Get_Struct(self, xmlNode, node);
-  Data_Get_Struct(rb_node, xmlNode, new_sibling);
-  xmlAddNextSibling(node, new_sibling);
+  Data_Get_Struct(rb_node, xmlNode, _new_sibling);
+
+  if(!(new_sibling = xmlAddNextSibling(node, _new_sibling)))
+    rb_raise(rb_eRuntimeError, "Could not add next sibling");
+
+  // the sibling was a text node that was coalesced. we need to have the object
+  // point at SOMETHING, or we'll totally bomb out.
+  if(new_sibling != _new_sibling) DATA_PTR(rb_node) = new_sibling;
 
   rb_funcall(rb_node, rb_intern("decorate!"), 0);
 
@@ -445,21 +459,46 @@ static VALUE add_next_sibling(VALUE self, VALUE rb_node)
  */
 static VALUE add_previous_sibling(VALUE self, VALUE rb_node)
 {
-  xmlNodePtr node, new_sibling;
+  xmlNodePtr node, sibling, new_sibling;
   Check_Type(rb_node, T_DATA);
 
   Data_Get_Struct(self, xmlNode, node);
-  Data_Get_Struct(rb_node, xmlNode, new_sibling);
+  Data_Get_Struct(rb_node, xmlNode, sibling);
 
-  if(!(new_sibling = xmlAddPrevSibling(node, new_sibling)))
+  if(!(new_sibling = xmlAddPrevSibling(node, sibling)))
     rb_raise(rb_eRuntimeError, "Could not add previous sibling");
 
+  // the sibling was a text node that was coalesced. we need to have the object
+  // point at SOMETHING, or we'll totally bomb out.
+  if(sibling != new_sibling) DATA_PTR(rb_node) = new_sibling;
+
   rb_funcall(rb_node, rb_intern("decorate!"), 0);
 
-  VALUE rb_new_sibling = Nokogiri_wrap_xml_node(new_sibling);
-  rb_funcall(rb_new_sibling, rb_intern("decorate!"), 0);
+  return rb_node;
+}
+
+/*
+ * call-seq:
+ *  to_html
+ *
+ * Returns this node as HTML
+ */
+static VALUE to_html(VALUE self)
+{
+  xmlBufferPtr buf ;
+  xmlNodePtr node ;
+  Data_Get_Struct(self, xmlNode, node);
+
+  VALUE html;
+
+  if(node->doc->type == XML_DOCUMENT_NODE)
+    return rb_funcall(self, rb_intern("to_xml"), 0);
 
-  return rb_new_sibling;
+  buf = xmlBufferCreate() ;
+  htmlNodeDump(buf, node->doc, node);
+  html = rb_str_new2((char*)buf->content);
+  xmlBufferFree(buf);
+  return html ;
 }
 
 /*
@@ -567,6 +606,10 @@ VALUE Nokogiri_wrap_xml_node(xmlNodePtr node)
       klass = rb_const_get(mNokogiriXml, rb_intern("Element"));
       rb_node = Data_Wrap_Struct(klass, 0, debug_node_dealloc, node) ;
       break;
+    case XML_ATTRIBUTE_NODE:
+      klass = rb_const_get(mNokogiriXml, rb_intern("Attr"));
+      rb_node = Data_Wrap_Struct(klass, 0, debug_node_dealloc, node) ;
+      break;
     case XML_ENTITY_DECL:
       klass = rb_const_get(mNokogiriXml, rb_intern("EntityDeclaration"));
       rb_node = Data_Wrap_Struct(klass, 0, debug_node_dealloc, node) ;
@@ -590,16 +633,12 @@ VALUE Nokogiri_wrap_xml_node(xmlNodePtr node)
 }
 
 
-void Nokogiri_xml_node_properties(xmlNodePtr node, VALUE attr_hash)
+void Nokogiri_xml_node_properties(xmlNodePtr node, VALUE attr_list)
 {
   xmlAttrPtr prop;
-  xmlChar* propstr ;
   prop = node->properties ;
   while (prop != NULL) {
-    propstr = xmlGetProp(node, prop->name) ;
-    rb_hash_aset(attr_hash, rb_str_new2((const char*)prop->name),
-                 rb_str_new2((char*)propstr));
-    xmlFree(propstr);
+    rb_ary_push(attr_list, Nokogiri_wrap_xml_node((xmlNodePtr)prop));
     prop = prop->next ;
   }
 }
@@ -645,9 +684,6 @@ void Nokogiri_xml_node_namespaces(xmlNodePtr node, VALUE attr_hash)
 VALUE cNokogiriXmlNode ;
 void init_xml_node()
 {
-  /*
-   * HACK.  This is so that rdoc will work with this C file.
-   */
   /*
   VALUE nokogiri = rb_define_module("Nokogiri");
   VALUE xml = rb_define_module_under(nokogiri, "XML");
@@ -672,13 +708,14 @@ void init_xml_node()
   rb_define_method(klass, "key?", key_eh, 1);
   rb_define_method(klass, "blank?", blank_eh, 0);
   rb_define_method(klass, "[]=", set, 2);
-  rb_define_method(klass, "remove_attribute", remove_prop, 1);
-  rb_define_method(klass, "attributes", attributes, 0);
+  rb_define_method(klass, "attribute_nodes", attribute_nodes, 0);
+  rb_define_method(klass, "namespace", namespace, 0);
   rb_define_method(klass, "namespaces", namespaces, 0);
   rb_define_method(klass, "add_previous_sibling", add_previous_sibling, 1);
   rb_define_method(klass, "add_next_sibling", add_next_sibling, 1);
   rb_define_method(klass, "encode_special_chars", encode_special_chars, 1);
   rb_define_method(klass, "to_xml", to_xml, -1);
+  rb_define_method(klass, "to_html", to_html, 0);
   rb_define_method(klass, "dup", duplicate_node, -1);
   rb_define_method(klass, "unlink", unlink_node, 0);
   rb_define_method(klass, "internal_subset", internal_subset, 0);