nokogiri 1.3.1 → 1.3.2

data/CHANGELOG.ja.rdoc

@@ -1,3 +1,28 @@
+=== 1.3.2 / ２００９年６月２２日
+
+* 新しい機能
+
+  * Nokogiri::XML::DTD#validate はドキュメントを検証できるようになった
+
+* バグの修正
+
+  * Nokogiri::XML::NodeSet#search はトップレベルのノードを検索するようになった
+    GH #73
+  * Nokogiri::XML::Documentからメソッドに関係する名前空間を取り除いた
+  * 同じ名前空間を２回目に入れた時に起るセグフォールトを修正した
+  * Snow Leopard で Nokogiri は動くようになった GH #79
+  * メーリングリストはGoogle Groupsの以下のアドレスに移動した
+    http://groups.google.com/group/nokogiri-talk
+  * HTML フラグメントはコメントとCDATAを正確に扱うようになった
+  * Nokogiri::XML::Document#cloneはdupのエイリアスと同様の働きをする
+
+* 廃棄予定
+
+  * Nokogiri::XML::SAX::Document#start_element_nsは廃棄予定なので
+    Nokogiri::XML::SAX::Document#start_element_namespaceを代わりに使用して下さい
+  * Nokogiri::XML::SAX::Document#end_element_nsは廃棄予定なので
+    Nokogiri::XML::SAX::Document#end_element_namespaceを代わりに使用して下さい
+
 === 1.3.1 / ２００９年６月７日
 
 * バグの修正

data/CHANGELOG.rdoc

@@ -1,3 +1,26 @@
+=== 1.3.2 / 2009-06-22
+
+* New Features
+
+  * Nokogiri::XML::DTD#validate will validate your document
+
+* Bugfixes
+
+  * Nokogiri::XML::NodeSet#search will search top level nodes. GH #73
+  * Removed namespace related methods from Nokogiri::XML::Document
+  * Fixed a segfault when a namespace was added twice
+  * Made nokogiri work with Snow Leopard GH #79
+  * Mailing list has moved to: http://groups.google.com/group/nokogiri-talk
+  * HTML fragments now correctly handle comments and CDATA blocks. GH #78
+  * Nokogiri::XML::Document#clone is now an alias of dup
+
+* Deprecations
+
+  * Nokogiri::XML::SAX::Document#start_element_ns is deprecated, please switch
+    to Nokogiri::XML::SAX::Document#start_element_namespace
+  * Nokogiri::XML::SAX::Document#end_element_ns is deprecated, please switch
+    to Nokogiri::XML::SAX::Document#end_element_namespace
+
 === 1.3.1 / 2009-06-07
 
 * Bugfixes

data/Manifest.txt

@@ -159,6 +159,7 @@ lib/nokogiri/xml/reader.rb
 lib/nokogiri/xml/relax_ng.rb
 lib/nokogiri/xml/sax.rb
 lib/nokogiri/xml/sax/document.rb
+lib/nokogiri/xml/sax/legacy_handlers.rb
 lib/nokogiri/xml/sax/parser.rb
 lib/nokogiri/xml/sax/push_parser.rb
 lib/nokogiri/xml/schema.rb
@@ -177,14 +178,18 @@ test/css/test_xpath_visitor.rb
 test/ffi/test_document.rb
 test/files/address_book.rlx
 test/files/address_book.xml
+test/files/bar/bar.xsd
 test/files/dont_hurt_em_why.xml
 test/files/exslt.xml
 test/files/exslt.xslt
+test/files/foo/foo.xsd
 test/files/po.xml
 test/files/po.xsd
+test/files/snuggles.xml
 test/files/staff.xml
 test/files/staff.xslt
 test/files/tlm.html
+test/files/valid_bar.xml
 test/helper.rb
 test/hpricot/files/basic.xhtml
 test/hpricot/files/boingboing.html

data/README.ja.rdoc

@@ -1,15 +1,15 @@
 = Nokogiri (鋸)
 
-* http://nokogiri.rubyforge.org/
+* http://nokogiri.org/
 * http://github.com/tenderlove/nokogiri/wikis
 * http://github.com/tenderlove/nokogiri/tree/master
-* http://rubyforge.org/mailman/listinfo/nokogiri-talk
+* http://groups.google.com/group/nokogiri-talk
 * http://github.com/tenderlove/nokogiri/issues
 
 == DESCRIPTION:
 
-Nokogiri はHTMLやXMLやSAXやXSLTやReaderのパーサーです。Nokogiri はドキュメント
-をXPath と CSS3 セレクター経由で検索する機能が特徴のひとつ。
+Nokogiri はHTMLとXMLとSAXとXSLTとReaderのパーサーです。とりわけ重要な特徴は、
+ドキュメントをXPathやCSS3セレクター経由で探索する機能を持つことです。
 
 XMLは暴力に似ている - XMLが君の問題を解決しないとしたら、君はXMLを十分に
 使いこなしていない事になる。
@@ -32,7 +32,7 @@ XML/HTMLの高速な解析と探索検索、ならびにCSS3セレクタとXPath
 
 Nokogiriのメーリングリスト:
 
-  * http://rubyforge.org/mailman/listinfo/nokogiri-talk
+  * http://groups.google.com/group/nokogiri-talk
 
 バグ報告:
 

data/README.rdoc

@@ -1,9 +1,9 @@
 = Nokogiri
 
-* http://nokogiri.rubyforge.org/
+* http://nokogiri.org/
 * http://github.com/tenderlove/nokogiri/wikis
 * http://github.com/tenderlove/nokogiri/tree/master
-* http://rubyforge.org/mailman/listinfo/nokogiri-talk
+* http://groups.google.com/group/nokogiri-talk
 * http://github.com/tenderlove/nokogiri/issues
 
 == DESCRIPTION:
@@ -35,7 +35,7 @@ to using correct CSS and XPath.
 
 The Nokogiri mailing list is available here:
 
-  * http://rubyforge.org/mailman/listinfo/nokogiri-talk
+  * http://groups.google.com/group/nokogiri-talk
 
 The bug tracker is available here:
 

data/Rakefile

@@ -1,42 +1,39 @@
 # -*- ruby -*-
 
 require 'rubygems'
-require 'rake'
+gem 'hoe', '>= 2.1.0'
 require 'hoe'
 
-LIB_DIR = File.expand_path(File.join(File.dirname(__FILE__), 'lib'))
-$LOAD_PATH << LIB_DIR
-
-windows = RUBY_PLATFORM =~ /(mswin|mingw)/i ? true : false
-java = RUBY_PLATFORM =~ /java/ ? true : false
+windows = RUBY_PLATFORM =~ /(mswin|mingw)/i
+java    = RUBY_PLATFORM =~ /java/
 
 GENERATED_PARSER    = "lib/nokogiri/css/generated_parser.rb"
 GENERATED_TOKENIZER = "lib/nokogiri/css/generated_tokenizer.rb"
 
-require 'nokogiri/version'
+# Make sure hoe-debugging is installed
+Hoe.plugin :debugging
 
-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 = [
+HOE = Hoe.spec 'nokogiri' do
+  developer('Aaron Patterson', 'aaronp@rubyforge.org')
+  developer('Mike Dalessio', 'mike.dalessio@gmail.com')
+  self.readme_file   = ['README', ENV['HLANG'], 'rdoc'].compact.join('.')
+  self.history_file  = ['CHANGELOG', ENV['HLANG'], 'rdoc'].compact.join('.')
+  self.extra_rdoc_files  = FileList['*.rdoc']
+  self.clean_globs = [
     'lib/nokogiri/*.{o,so,bundle,a,log,dll}',
     GENERATED_PARSER,
     GENERATED_TOKENIZER,
     'cross',
   ]
 
-  p.extra_dev_deps  << "racc"
-  p.extra_dev_deps  << "rexical"
-  p.extra_dev_deps  << "rake-compiler"
+  %w{ racc rexical rake-compiler }.each do |dep|
+    extra_dev_deps << dep
+  end
 
-  p.spec_extras = { :extensions => ["ext/nokogiri/extconf.rb"] }
+  self.spec_extras = { :extensions => ["ext/nokogiri/extconf.rb"] }
 end
 
 unless java
-
   gem 'rake-compiler', '>= 0.4.1'
   require "rake/extensiontask"
 
@@ -46,7 +43,7 @@ unless java
     ext.config_options << ENV['EXTOPTS']
     cross_dir = File.join(File.dirname(__FILE__), 'tmp', 'cross')
     ext.cross_compile   = true
-    ext.cross_platform  = 'i386-mswin32'
+    ext.cross_platform  = 'i386-mingw32'
     ext.cross_config_options <<
       "--with-iconv-dir=#{File.join(cross_dir, 'iconv')}"
     ext.cross_config_options <<
@@ -74,12 +71,12 @@ require "#{HOE.name}/\#{RUBY_VERSION.sub(/\\.\\d+$/, '')}/#{HOE.name}"
       eoruby
     end
     HOE.spec.extensions = []
-    HOE.spec.platform = 'x86-mswin32'
+    HOE.spec.platform = 'x86-mingw32'
     HOE.spec.files += Dir["lib/#{HOE.name}/#{HOE.name}.rb"]
     HOE.spec.files += Dir["lib/#{HOE.name}/1.{8,9}/*"]
     HOE.spec.files += Dir["ext/nokogiri/*.dll"]
   end
-  CLOBBER.include("lib/nokogiri/nokogiri.rb")
+  CLOBBER.include("lib/nokogiri/nokogiri.{so,dylib,rb,bundle}")
   CLOBBER.include("lib/nokogiri/1.{8,9}")
 end
 
@@ -184,9 +181,16 @@ unless windows || java || ENV['NOKOGIRI_FFI']
   end
 
   Rake::Task[:test].prerequisites << :compile
-  ['valgrind', 'valgrind_mem', 'valgrind_mem0', 'coverage'].each do |task_name|
+  ['valgrind', 'valgrind:mem', 'valgrind:mem0'].each do |task_name|
     Rake::Task["test:#{task_name}"].prerequisites << :compile
   end
+else
+  [:test, :check_manifest].each do |task_name|
+    if Rake::Task[task_name]
+      Rake::Task[task_name].prerequisites << GENERATED_PARSER
+      Rake::Task[task_name].prerequisites << GENERATED_TOKENIZER
+    end
+  end
 end
 
 namespace :install do

data/ext/nokogiri/extconf.rb

@@ -1,4 +1,4 @@
-ENV["ARCHFLAGS"] = "-arch #{`uname -p` =~ /powerpc/ ? 'ppc' : 'i386'}"
+ENV['RC_ARCHS'] = '' if RUBY_PLATFORM =~ /darwin/
 
 require 'mkmf'
 
@@ -54,35 +54,60 @@ LIB_DIRS = [
 ]
 
 iconv_dirs = dir_config('iconv', '/opt/local/include', '/opt/local/lib')
-unless [nil, nil] == iconv_dirs
+unless ["", ""] == iconv_dirs
   HEADER_DIRS.unshift iconv_dirs.first
   LIB_DIRS.unshift iconv_dirs[1]
 end
 
 xml2_dirs = dir_config('xml2', '/opt/local/include/libxml2', '/opt/local/lib')
-unless [nil, nil] == xml2_dirs
+unless ["", ""] == xml2_dirs
   HEADER_DIRS.unshift xml2_dirs.first
   LIB_DIRS.unshift xml2_dirs[1]
 end
 
 xslt_dirs = dir_config('xslt', '/opt/local/include/', '/opt/local/lib')
-unless [nil, nil] == xslt_dirs
+unless ["", ""] == xslt_dirs
   HEADER_DIRS.unshift xslt_dirs.first
   LIB_DIRS.unshift xslt_dirs[1]
 end
 
-unless find_header('iconv.h', *HEADER_DIRS)
+CUSTOM_DASH_I = []
+
+def nokogiri_find_header header_file, *paths
+  message = checking_message(header_file, paths)
+  header = cpp_include header_file
+  checking_for message do
+    found = false
+    paths.each do |dir|
+      if File.exists?(File.join(dir, header_file))
+        opt = "-I#{dir}".quote
+        if try_cpp header, opt
+          unless CUSTOM_DASH_I.include? dir
+            $INCFLAGS = "#{opt} #{$INCFLAGS}"
+            CUSTOM_DASH_I << dir
+          end
+          found = dir
+          break
+        end
+      end
+    end
+    found ||= try_cpp(header)
+  end
+end
+
+unless nokogiri_find_header('iconv.h', *HEADER_DIRS)
   abort "iconv is missing.  try 'port install iconv' or 'yum install iconv'"
 end
 
-unless find_header('libxml/parser.h', *HEADER_DIRS)
+unless nokogiri_find_header('libxml/parser.h', *HEADER_DIRS)
   abort "libxml2 is missing.  try 'port install libxml2' or 'yum install libxml2-devel'"
 end
 
-unless find_header('libxslt/xslt.h', *HEADER_DIRS)
+unless nokogiri_find_header('libxslt/xslt.h', *HEADER_DIRS)
   abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt-devel'"
 end
-unless find_header('libexslt/exslt.h', *HEADER_DIRS)
+
+unless nokogiri_find_header('libexslt/exslt.h', *HEADER_DIRS)
   abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt-devel'"
 end
 
@@ -98,10 +123,27 @@ unless find_library('exslt', 'exsltFuncRegister', *LIB_DIRS)
   abort "libxslt is missing.  try 'port install libxslt' or 'yum install libxslt-devel'"
 end
 
-have_func('xmlRelaxNGSetParserStructuredErrors')
-have_func('xmlRelaxNGSetValidStructuredErrors')
-have_func('xmlSchemaSetValidStructuredErrors')
-have_func('xmlSchemaSetParserStructuredErrors')
+def nokogiri_link_command ldflags, opt='', libpath=$LIBPATH
+  old_link_command ldflags, opt, libpath
+end
+
+def with_custom_link
+  alias :old_link_command :link_command
+  alias :link_command :nokogiri_link_command
+  yield
+ensure
+  alias :link_command :old_link_command
+end
+
+with_custom_link do
+  with_cppflags $INCFLAGS do
+    have_func('xmlRelaxNGSetParserStructuredErrors')
+    have_func('xmlRelaxNGSetParserStructuredErrors')
+    have_func('xmlRelaxNGSetValidStructuredErrors')
+    have_func('xmlSchemaSetValidStructuredErrors')
+    have_func('xmlSchemaSetParserStructuredErrors')
+  end
+end
 
 if ENV['CPUPROFILE']
   unless find_library('profiler', 'ProfilerEnable', *LIB_DIRS)

data/ext/nokogiri/xml_document.c

@@ -296,12 +296,15 @@ VALUE Nokogiri_wrap_xml_document(VALUE klass, xmlDocPtr doc)
       dealloc,
       doc
   );
+
+  VALUE cache = rb_ary_new();
   rb_iv_set(rb_doc, "@decorators", Qnil);
-  rb_iv_set(rb_doc, "@node_cache", rb_ary_new());
+  rb_iv_set(rb_doc, "@node_cache", cache);
   rb_funcall(rb_doc, rb_intern("initialize"), 0);
 
   tuple->doc = (void *)rb_doc;
   tuple->unlinkedNodes = xmlXPathNodeSetCreate(NULL);
+  tuple->node_cache = cache;
   doc->_private = tuple ;
 
   return rb_doc ;

data/ext/nokogiri/xml_document.h

@@ -6,6 +6,7 @@
 struct _nokogiriTuple {
   xmlDocPtr     doc;
   xmlNodeSetPtr unlinkedNodes;
+  VALUE         node_cache;
 };
 typedef struct _nokogiriTuple nokogiriTuple;
 typedef nokogiriTuple * nokogiriTuplePtr;
@@ -16,6 +17,7 @@ VALUE Nokogiri_wrap_xml_document(VALUE klass, xmlDocPtr doc);
 #define DOC_RUBY_OBJECT_TEST(x) ((nokogiriTuplePtr)(x->_private))
 #define DOC_RUBY_OBJECT(x) ((VALUE)((nokogiriTuplePtr)(x->_private))->doc)
 #define DOC_UNLINKED_NODE_SET(x) ((xmlNodeSetPtr)((nokogiriTuplePtr)(x->_private))->unlinkedNodes)
+#define DOC_NODE_CACHE(x) ((VALUE)((nokogiriTuplePtr)(x->_private))->node_cache)
 
 extern VALUE cNokogiriXmlDocument ;
 #endif

data/ext/nokogiri/xml_dtd.c

@@ -85,6 +85,34 @@ static VALUE elements(VALUE self)
   return hash;
 }
 
+/*
+ * call-seq:
+ *   validate(document)
+ *
+ * Validate +document+ returning a list of errors
+ */
+static VALUE validate(VALUE self, VALUE document)
+{
+  xmlDocPtr doc;
+  xmlDtdPtr dtd;
+
+  Data_Get_Struct(self, xmlDtd, dtd);
+  Data_Get_Struct(document, xmlDoc, doc);
+  VALUE error_list      = rb_ary_new();
+
+  xmlValidCtxtPtr ctxt = xmlNewValidCtxt();
+
+  xmlSetStructuredErrorFunc((void *)error_list, Nokogiri_error_array_pusher);
+
+  xmlValidateDtd(ctxt, doc, dtd);
+
+  xmlSetStructuredErrorFunc(NULL, NULL);
+
+  xmlFreeValidCtxt(ctxt);
+
+  return error_list;
+}
+
 void init_xml_dtd()
 {
   VALUE nokogiri = rb_define_module("Nokogiri");
@@ -99,4 +127,5 @@ void init_xml_dtd()
   rb_define_method(klass, "notations", notations, 0);
   rb_define_method(klass, "elements", elements, 0);
   rb_define_method(klass, "entities", entities, 0);
+  rb_define_method(klass, "validate", validate, 1);
 }

data/ext/nokogiri/xml_node.c

@@ -720,6 +720,14 @@ static VALUE add_namespace_definition(VALUE self, VALUE prefix, VALUE href)
       (const xmlChar *)(prefix == Qnil ? NULL : StringValuePtr(prefix))
   );
 
+  if(!ns) {
+    ns = xmlSearchNs(
+        node->doc,
+        node,
+        (const xmlChar *)(prefix == Qnil ? NULL : StringValuePtr(prefix))
+    );
+  }
+
   if(Qnil == prefix) xmlSetNs(node, ns);
 
   return Nokogiri_wrap_xml_namespace(node->doc, ns);
@@ -849,7 +857,7 @@ VALUE Nokogiri_wrap_xml_node(VALUE klass, xmlNodePtr node)
 
   if (DOC_RUBY_OBJECT_TEST(node->doc) && DOC_RUBY_OBJECT(node->doc)) {
     document = DOC_RUBY_OBJECT(node->doc);
-    node_cache = rb_iv_get(document, "@node_cache");
+    node_cache = DOC_NODE_CACHE(node->doc);
   }
 
   rb_ary_push(node_cache, rb_node);

data/ext/nokogiri/xml_node_set.c

@@ -145,7 +145,11 @@ static VALUE plus(VALUE self, VALUE rb_other)
   new = xmlXPathNodeSetMerge(NULL, node_set);
   new = xmlXPathNodeSetMerge(new, other);
 
-  return Nokogiri_wrap_xml_node_set(new);
+  VALUE new_set = Nokogiri_wrap_xml_node_set(new);
+
+  rb_iv_set(new_set, "@document", rb_iv_get(self, "@document"));
+
+  return new_set;
 }
 
 /*