nokogiri 1.2.1 → 1.2.2

data/.autotest

@@ -0,0 +1,15 @@
+# -*- ruby -*-
+
+Autotest.add_hook :run_command do |at|
+  if ENV['ONENINE']
+    system "rake1.9 build"
+  else
+    system "rake build"
+  end
+end
+
+class Autotest
+  def ruby
+    'ruby1.9'
+  end
+end if ENV['ONENINE']

data/{History.ja.txt → CHANGELOG.ja.rdoc}

@@ -1,11 +1,39 @@
-=== 1.2.1 / ２００８年２月２３日
+=== 1.2.2 / ２００９年３月１４日
+
+* 新しい機能
+
+  * Nokogiri は soap4r と一緒に使う事が可能。(XSD::XMLParser::Nokogiri 参照)
+  * Nokogiri::XML::Node#inner_html= はノードの中のHTMLをセット出来る
+  * Nokogiriのビルダーのインタフェースの改良
+  * Nokogiri::XML::Node#swap　は、現在のノードに新しいhtmlを交換する事が出来る
+
+
+* バグの修正
+
+  * ビルダーAPIのタグのネスティングを修正 (LH #41)
+  * Nokogiri::HTML.fragment はテキストだけのノードを適切に扱う事が出来る(LH #43)
+  * Nokogiri::XML::Node#before はテキストノードのはじめに挿入する事が出来る (LH #44)
+  * Nokogiri::XML::Node#after はテキストノードの文末に挿入する事が出来る
+  * Nokogiri::XML::Node#search 名前空間が自動的に登録されるようになった(LH#42)
+  * Nokogiri::XML::NodeSet#search 名前空間が自動的に登録されるようになった
+  * Nokogiri::HTML::NamedCharacters はlibxml2に委任
+  * Nokogiri::XML::Node#[] はSymbolを使う (LH #48)
+  * vasprintf にwindowsを修正 (Geffroy Couprie ありがとう！)
+  * Nokogiri::XML::Node#[]= はentityを符号化しない (LH #55)
+  * 名前空間はreparentedのノードに模写する (LH #56)
+  * StringのエンコードはRuby 1.9での初期設定を使用する
+  * Document#dup は新しいドキュメントに同じタイプを作る (LH #59)
+  * Document#parent は存在しない (LH #64)
+
+
+=== 1.2.1 / ２００９年２月２３日
 
 * 修正
 
   * CSS のセレクターのスペースを修正
   * Ruby 1.9 のStringのエンコードを修正 （角谷さんに感謝！）
 
-=== 1.2.0 / ２００８年２月２２日
+=== 1.2.0 / ２００９年２月２２日
 
 * 新しい機能
   * CSSサーチが CSS3 名前空間クエリをサポートするようになった

data/{History.txt → CHANGELOG.rdoc}

@@ -1,11 +1,37 @@
-=== 1.2.1 / 2008-02-23
+=== 1.2.2 / 2009-03-14
+
+* New features
+
+  * Nokogiri may be used with soap4r. See XSD::XMLParser::Nokogiri
+  * Nokogiri::XML::Node#inner_html= to set the inner html for a node
+  * Nokogiri builder interface improvements
+  * Nokogiri::XML::Node#swap swaps html for current node (LH #50)
+
+* Bugfixes
+
+  * Fixed a tag nesting problem in the Builder API (LH #41)
+  * Nokogiri::HTML.fragment will properly handle text only nodes (LH #43)
+  * Nokogiri::XML::Node#before will prepend text nodes (LH #44)
+  * Nokogiri::XML::Node#after will append text nodes
+  * Nokogiri::XML::Node#search automatically registers root namepsaces (LH #42)
+  * Nokogiri::XML::NodeSet#search automatically registers namespaces
+  * Nokogiri::HTML::NamedCharacters delegates to libxml2
+  * Nokogiri::XML::Node#[] can take a symbol (LH #48)
+  * vasprintf for windows updated.  Thanks Geoffroy Couprie!
+  * Nokogiri::XML::Node#[]= should not encode entities (LH #55)
+  * Namespaces should be copied to reparented nodes (LH #56)
+  * Nokogiri uses encoding set on the string for default in Ruby 1.9
+  * Document#dup should create a new document of the same type (LH #59)
+  * Document should not have a parent method (LH #64)
+
+=== 1.2.1 / 2009-02-23
 
 * Bugfixes
 
   * Fixed a CSS selector space bug
   * Fixed Ruby 1.9 String Encoding (Thanks 角谷さん！)
 
-=== 1.2.0 / 2008-02-22
+=== 1.2.0 / 2009-02-22
 
 * New features
 

data/Manifest.txt

@@ -1,12 +1,15 @@
-History.ja.txt
-History.txt
+.autotest
+CHANGELOG.ja.rdoc
+CHANGELOG.rdoc
 Manifest.txt
-README.ja.txt
-README.txt
+README.ja.rdoc
+README.rdoc
 Rakefile
 ext/nokogiri/extconf.rb
 ext/nokogiri/html_document.c
 ext/nokogiri/html_document.h
+ext/nokogiri/html_entity_lookup.c
+ext/nokogiri/html_entity_lookup.h
 ext/nokogiri/html_sax_parser.c
 ext/nokogiri/html_sax_parser.h
 ext/nokogiri/native.c
@@ -71,13 +74,12 @@ lib/nokogiri/hpricot.rb
 lib/nokogiri/html.rb
 lib/nokogiri/html/builder.rb
 lib/nokogiri/html/document.rb
+lib/nokogiri/html/entity_lookup.rb
 lib/nokogiri/html/sax/parser.rb
 lib/nokogiri/syntax_error.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
 lib/nokogiri/xml/comment.rb
@@ -86,6 +88,7 @@ lib/nokogiri/xml/document_fragment.rb
 lib/nokogiri/xml/dtd.rb
 lib/nokogiri/xml/element.rb
 lib/nokogiri/xml/entity_declaration.rb
+lib/nokogiri/xml/fragment_handler.rb
 lib/nokogiri/xml/node.rb
 lib/nokogiri/xml/node/save_options.rb
 lib/nokogiri/xml/node_set.rb
@@ -103,6 +106,7 @@ lib/nokogiri/xml/xpath/syntax_error.rb
 lib/nokogiri/xml/xpath_context.rb
 lib/nokogiri/xslt.rb
 lib/nokogiri/xslt/stylesheet.rb
+lib/xsd/xmlparser/nokogiri.rb
 tasks/test.rb
 test/css/test_nthiness.rb
 test/css/test_parser.rb
@@ -135,6 +139,8 @@ test/hpricot/test_xml.rb
 test/html/sax/test_parser.rb
 test/html/test_builder.rb
 test/html/test_document.rb
+test/html/test_document_encoding.rb
+test/html/test_named_characters.rb
 test/html/test_node.rb
 test/test_convert_xpath.rb
 test/test_css_cache.rb
@@ -162,5 +168,5 @@ test/xml/test_node_set.rb
 test/xml/test_processing_instruction.rb
 test/xml/test_reader_encoding.rb
 test/xml/test_text.rb
+test/xml/test_unparented_node.rb
 test/xml/test_xpath.rb
-vendor/hoe.rb

data/{README.ja.txt → README.ja.rdoc}

@@ -64,8 +64,10 @@ NokogiriはHpricotの代わりに使用出来る。
 == REQUIREMENTS:
 
 * ruby 1.8 or 1.9
-* libxml
+* libxml2
+* libxml2-dev
 * libxslt
+* libxslt-dev
 
 == INSTALL:
 

data/{README.txt → README.rdoc}

@@ -42,8 +42,12 @@ The bug tracker is available here:
   require 'nokogiri'
   require 'open-uri'
   
+  # Get a Nokogiri::HTML:Document for the page we’re interested in...
+
   doc = Nokogiri::HTML(open('http://www.google.com/search?q=tenderlove'))
   
+  # Do funky things with it using Nokogiri::XML::Node methods...
+
   ####
   # Search for nodes by css
   doc.css('h3.r a.l').each do |link|
@@ -66,8 +70,10 @@ The bug tracker is available here:
 == REQUIREMENTS:
 
 * ruby 1.8 or 1.9
-* libxml
+* libxml2
+* libxml2-dev
 * libxslt
+* libxslt-dev
 
 == INSTALL:
 

data/Rakefile

@@ -2,6 +2,7 @@
 
 require 'rubygems'
 require 'rake'
+require 'hoe'
 
 
 kind = Config::CONFIG['DLEXT']
@@ -10,8 +11,6 @@ windows = RUBY_PLATFORM =~ /mswin/i ? true : false
 LIB_DIR = File.expand_path(File.join(File.dirname(__FILE__), 'lib'))
 $LOAD_PATH << LIB_DIR
 
-require 'vendor/hoe'
-
 GENERATED_PARSER = "lib/nokogiri/css/generated_parser.rb"
 GENERATED_TOKENIZER = "lib/nokogiri/css/generated_tokenizer.rb"
 
@@ -22,6 +21,9 @@ require 'nokogiri/version'
 HOE = Hoe.new('nokogiri', Nokogiri::VERSION) do |p|
   p.developer('Aaron Patterson', 'aaronp@rubyforge.org')
   p.developer('Mike Dalessio', 'mike.dalessio@gmail.com')
+  p.readme_file   = ['README', ENV['HLANG'], 'rdoc'].compact.join('.')
+  p.history_file  = ['CHANGELOG', ENV['HLANG'], 'rdoc'].compact.join('.')
+  p.extra_rdoc_files  = FileList['*.rdoc']
   p.clean_globs = [
     'ext/nokogiri/Makefile',
     'ext/nokogiri/*.{o,so,bundle,a,log,dll}',
@@ -102,7 +104,7 @@ file GENERATED_TOKENIZER => "lib/nokogiri/css/tokenizer.rex" do |t|
   begin
     sh "frex --independent -o #{t.name} #{t.prerequisites.first}"
   rescue
-    abort "need frex, sudo gem install aaronp-frex -s http://gems.github.com"   
+    abort "need frex, sudo gem install tenderlove-frex -s http://gems.github.com"   
   end
 end
 
@@ -254,31 +256,12 @@ namespace :install do
   desc "Install frex and racc for development"
   task :deps => %w(frex racc)
 
-  directory "stash"
-
-  file "stash/racc-1.4.5-all.tar.gz" => "stash" do |t|
-    puts "Downloading racc to #{t.name}..."
-
-    Dir.chdir File.dirname(t.name) do
-      url = "http://i.loveruby.net/archive/racc/racc-1.4.5-all.tar.gz"
-      system "wget #{url} || curl -O #{url}"
-    end
-  end
-
-  task :racc => "stash/racc-1.4.5-all.tar.gz" do |t|
-    sh "tar xvf #{t.prerequisites.first} -C stash"
-
-    Dir.chdir "stash/#{File.basename(t.prerequisites.first, ".tar.gz")}" do
-      sh "ruby setup.rb config"
-      sh "ruby setup.rb setup"
-      sh "sudo ruby setup.rb install"
-    end
-
-    puts "The racc binary is likely in #{::Config::CONFIG["bindir"]}."
+  task :racc do |t|
+    sh "sudo gem install racc"
   end
 
   task :frex do
-    sh "sudo gem install aaronp-frex -s http://gems.github.com"
+    sh "sudo gem install tenderlove-frex -s http://gems.github.com"
   end
 end
 

data/ext/nokogiri/extconf.rb

@@ -57,10 +57,10 @@ unless find_header('libxml/parser.h', *HEADER_DIRS)
 end
 
 unless find_header('libxslt/xslt.h', *HEADER_DIRS)
-  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt'"
+  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt-devel'"
 end
 unless find_header('libexslt/exslt.h', *HEADER_DIRS)
-  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt'"
+  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt-devel'"
 end
 
 unless find_library('xml2', 'xmlParseDoc', *LIB_DIRS)
@@ -68,11 +68,11 @@ unless find_library('xml2', 'xmlParseDoc', *LIB_DIRS)
 end
 
 unless find_library('xslt', 'xsltParseStylesheetDoc', *LIB_DIRS)
-  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt'"
+  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt-devel'"
 end
 
 unless find_library('exslt', 'exsltFuncRegister', *LIB_DIRS)
-  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt'"
+  abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt-devel'"
 end
 
 create_makefile('nokogiri/native')

data/ext/nokogiri/html_entity_lookup.c

@@ -0,0 +1,30 @@
+#include <html_entity_lookup.h>
+
+/*
+ * call-seq:
+ *  get(key)
+ *
+ * Get the HTML::EntityDescription for +key+
+ */
+static VALUE get(VALUE self, VALUE key)
+{
+  const htmlEntityDesc * desc =
+    htmlEntityLookup((const xmlChar *)StringValuePtr(key));
+
+  if(NULL == desc) return Qnil;
+  VALUE klass = rb_const_get(mNokogiriHtml, rb_intern("EntityDescription"));
+
+  return rb_funcall(klass, rb_intern("new"), 3,
+      INT2NUM((int)desc->value),
+      NOKOGIRI_STR_NEW2(desc->name, "UTF-8"),
+      NOKOGIRI_STR_NEW2(desc->desc, "UTF-8"));
+}
+
+void init_html_entity_lookup()
+{
+  VALUE nokogiri = rb_define_module("Nokogiri");
+  VALUE html = rb_define_module_under(nokogiri, "HTML");
+  VALUE klass = rb_define_class_under(html, "EntityLookup", rb_cObject);
+
+  rb_define_method(klass, "get", get, 1);
+}

data/ext/nokogiri/html_entity_lookup.h

@@ -0,0 +1,8 @@
+#ifndef NOKOGIRI_HTML_ENTITY_LOOKUP
+#define NOKOGIRI_HTML_ENTITY_LOOKUP
+
+#include <native.h>
+
+void init_html_entity_lookup();
+
+#endif

data/ext/nokogiri/native.c

@@ -7,6 +7,23 @@ VALUE mNokogiriXslt ;
 VALUE mNokogiriXmlSax ;
 VALUE mNokogiriHtmlSax ;
 
+#ifdef XP_WIN
+/*
+ * I srsly hate windows.  it doesn't have vasprintf.
+ * Thank you Geoffroy Couprie for this implementation of vasprintf!
+ */
+int vasprintf (char **strp, const char *fmt, va_list ap)
+{
+  int len = vsnprintf (NULL, 0, fmt, ap) + 1;
+  char *res = (char *)malloc((unsigned int)len);
+  if (res == NULL)
+      return -1;
+  *strp = res;
+  return vsnprintf(res, (unsigned int)len, fmt, ap);
+}
+#endif
+
+
 void Init_native()
 {
   xmlMemSetup(
@@ -27,6 +44,10 @@ void Init_native()
                 rb_intern("LIBXML_VERSION"),
                 NOKOGIRI_STR_NEW2(LIBXML_DOTTED_VERSION, "UTF-8")
               );
+  rb_const_set( mNokogiri,
+                rb_intern("LIBXML_PARSER_VERSION"),
+                NOKOGIRI_STR_NEW2(xmlParserVersion, "UTF-8")
+              );
 
   init_xml_document();
   init_html_document();
@@ -48,4 +69,5 @@ void Init_native()
   init_html_sax_parser();
   init_xslt_stylesheet();
   init_xml_syntax_error();
+  init_html_entity_lookup();
 }

data/ext/nokogiri/native.h

@@ -30,16 +30,26 @@
 #define NOKOGIRI_STR_NEW2(str, encoding) \
   ({ \
     VALUE _string = rb_str_new2((const char *)str); \
-    if(NULL != encoding) \
-      rb_enc_associate_index(_string, rb_enc_find_index(encoding)); \
+    if(NULL != encoding) { \
+      int _enc = rb_enc_find_index(encoding); \
+      if(_enc == -1) \
+        rb_enc_associate_index(_string, rb_enc_find_index("ASCII")); \
+      else \
+        rb_enc_associate_index(_string, _enc); \
+    } \
     _string; \
   })
 
 #define NOKOGIRI_STR_NEW(str, len, encoding) \
   ({ \
     VALUE _string = rb_str_new((const char *)str, (long)len); \
-    if(NULL != encoding) \
-      rb_enc_associate_index(_string, rb_enc_find_index(encoding)); \
+    if(NULL != encoding) { \
+      int _enc = rb_enc_find_index(encoding); \
+      if(_enc == -1) \
+        rb_enc_associate_index(_string, rb_enc_find_index("ASCII")); \
+      else \
+        rb_enc_associate_index(_string, _enc); \
+    } \
     _string; \
   })
 
@@ -52,8 +62,13 @@
   rb_str_new((const char *)str, (long)len)
 #endif
 
+#ifdef XP_WIN
+int vasprintf (char **strp, const char *fmt, va_list ap);
+#endif
+
 #include <xml_io.h>
 #include <xml_document.h>
+#include <html_entity_lookup.h>
 #include <html_document.h>
 #include <xml_node.h>
 #include <xml_text.h>
@@ -81,6 +96,14 @@ extern VALUE mNokogiriHtml ;
 extern VALUE mNokogiriHtmlSax ;
 extern VALUE mNokogiriXslt ;
 
+#define NOKOGIRI_ROOT_NODE(_node) \
+  ({ \
+    nokogiriTuplePtr tuple = (nokogiriTuplePtr)(_node->doc->_private);       \
+    xmlNodeSetPtr node_set = (xmlNodeSetPtr)(tuple->unlinkedNodes);     \
+    xmlXPathNodeSetAdd(node_set, _node); \
+    _node; \
+  })
+
 #ifdef DEBUG
 
 #define NOKOGIRI_DEBUG_START(p) if (getenv("NOKOGIRI_NO_FREE")) return ; if (getenv("NOKOGIRI_DEBUG")) fprintf(stderr,"nokogiri: %s:%d %p start\n", __FILE__, __LINE__, p);

data/ext/nokogiri/xml_document.c

@@ -3,8 +3,23 @@
 static void dealloc(xmlDocPtr doc)
 {
   NOKOGIRI_DEBUG_START(doc);
+
+  nokogiriTuplePtr tuple = doc->_private;
+  xmlNodeSetPtr node_set = tuple->unlinkedNodes;
+
+  int i;
+  for(i = 0; i < node_set->nodeNr; i++) {
+    xmlAddChild((xmlNodePtr)doc, node_set->nodeTab[i]);
+  }
+
+  if (node_set->nodeTab != NULL)
+    xmlFree(node_set->nodeTab);
+  xmlFree(node_set);
+
+  free(doc->_private);
   doc->_private = NULL;
   xmlFreeDoc(doc);
+
   NOKOGIRI_DEBUG_END(doc);
 }
 
@@ -187,7 +202,11 @@ static VALUE duplicate_node(int argc, VALUE *argv, VALUE self)
   dup = xmlCopyDoc(doc, NUM2INT(level));
   if(dup == NULL) return Qnil;
 
-  return Nokogiri_wrap_xml_document(cNokogiriXmlDocument, dup);
+  dup->type = doc->type;
+  if(dup->type == XML_DOCUMENT_NODE)
+    return Nokogiri_wrap_xml_document(cNokogiriXmlDocument, dup);
+  else
+    return Nokogiri_wrap_xml_document(cNokogiriHtmlDocument, dup);
 }
 
 /*
@@ -255,7 +274,6 @@ void init_xml_document()
   rb_define_method(klass, "encoding", encoding, 0);
   rb_define_method(klass, "dup", duplicate_node, -1);
   rb_define_method(klass, "url", url, 0);
-  rb_undef_method(klass, "parent");
 }
 
 
@@ -263,10 +281,19 @@ void init_xml_document()
 VALUE Nokogiri_wrap_xml_document(VALUE klass, xmlDocPtr doc)
 {
   VALUE rb_doc = Qnil;
+  nokogiriTuplePtr tuple = (nokogiriTuplePtr)malloc(sizeof(nokogiriTuple));
 
-  rb_doc = Data_Wrap_Struct(klass ? klass : cNokogiriXmlDocument, 0, dealloc, doc) ;
+  rb_doc = Data_Wrap_Struct(
+      klass ? klass : cNokogiriXmlDocument,
+      0,
+      dealloc,
+      doc
+  );
   rb_iv_set(rb_doc, "@decorators", Qnil);
-  doc->_private = (void *)rb_doc;
+
+  tuple->doc = (void *)rb_doc;
+  tuple->unlinkedNodes = xmlXPathNodeSetCreate(NULL);
+  doc->_private = tuple ;
 
   return rb_doc ;
 }