nokogiri 1.0.7 → 1.1.0

data/History.ja.txt

@@ -1,3 +1,17 @@
+=== HEAD
+
+* 新しい機能
+
+  * 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
+
+  * 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
+
 === 1.0.7
 
 * バグの修正

data/History.txt

@@ -1,3 +1,19 @@
+=== 1.1.0
+
+* New Features
+
+  * Custom XPath functions are now supported.  See Nokogiri::XML::Node#xpath
+  * Custom CSS pseudo classes are now supported.  See Nokogiri::XML::Node#css
+  * Nokogiri::XML::Node#<< will add a child to the current node
+
+* Bugfixes
+
+  * Mutex lock on CSS cache access
+  * Fixed build problems with GCC 3.3.5
+  * XML::Node#to_xml now takes an indentation argument
+  * XML::Node#dup takes an optional depth argument
+  * XML::Node#add_previous_sibling returns new sibling node.
+
 === 1.0.7
 
 * Bugfixes

data/Manifest.txt

@@ -47,6 +47,7 @@ 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
@@ -85,6 +86,7 @@ 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
@@ -92,6 +94,8 @@ test/css/test_parser.rb
 test/css/test_tokenizer.rb
 test/css/test_xpath_visitor.rb
 test/files/dont_hurt_em_why.xml
+test/files/exslt.xml
+test/files/exslt.xslt
 test/files/staff.xml
 test/files/staff.xslt
 test/files/tlm.html
@@ -132,4 +136,5 @@ test/xml/test_dtd.rb
 test/xml/test_node.rb
 test/xml/test_node_set.rb
 test/xml/test_text.rb
+test/xml/test_xpath.rb
 vendor/hoe.rb

data/README.ja.txt

@@ -20,7 +20,7 @@ Nokogiri はHTMLやXMLやSAXやXSLTやReaderのパーサーです。
 
 これはスピードテストです：
 
-  * http://gist.github.com/22176
+  * http://gist.github.com/24605
 
 NokogiriはHpricotの代わりに使用出来る。
 その互換性は簡単に正しいCSSとXPathを使用する事が出来る。

data/README.txt

@@ -22,7 +22,7 @@ correctly implemented CSS3 selector support as well as XPath support.
 
 Here is a speed test:
 
-  * http://gist.github.com/22176
+  * http://gist.github.com/24605
 
 Nokogiri also features an Hpricot compatibility layer to help ease the change
 to using correct CSS and XPath.

data/Rakefile

@@ -231,9 +231,13 @@ end
 
 def test_suite_cmdline
   require 'find'
+  match = ENV['MATCH'] ? /#{ENV['MATCH']}/ : /./
+
   files = []
   Find.find("test") do |f|
-    files << f if File.basename(f) =~ /.*test.*\.rb$/
+    basename = File.basename(f)
+
+    files << f if basename =~ /.*test.*\.rb$/ && basename =~ match
   end
   cmdline = "ruby -w -I.:lib:ext:test -rtest/unit -e '%w[#{files.join(' ')}].each {|f| require f}'"
 end
@@ -328,11 +332,4 @@ unless windows
   Rake::Task[:check_manifest].prerequisites << GENERATED_TOKENIZER
 end
 
-# Evil evil hack.  Do not run tests when gem installs
-if ENV['RUBYARCHDIR']
-  prereqs = Rake::Task[:default].prerequisites
-  prereqs.clear
-  prereqs << :build
-end
-
 # vim: syntax=Ruby

data/ext/nokogiri/extconf.rb

@@ -6,6 +6,8 @@ ROOT = File.expand_path(File.join(File.dirname(__FILE__), '..', '..'))
 LIBDIR = Config::CONFIG['libdir']
 INCLUDEDIR = Config::CONFIG['includedir']
 
+  use_macports = !(defined?(RUBY_ENGINE) && RUBY_ENGINE != 'ruby')
+
 $CFLAGS << " #{ENV["CFLAGS"]}"
 if Config::CONFIG['target_os'] == 'mingw32'
   $CFLAGS << " -DXP_WIN -DXP_WIN32"
@@ -13,16 +15,36 @@ else
   $CFLAGS << " -g -DXP_UNIX"
 end
 
-$CFLAGS << " -O3 -Wall -Wextra -Wcast-qual -Wwrite-strings -Wconversion -Wmissing-noreturn -Winline"
+$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)
-  find_library('xslt', 'xsltParseStylesheetDoc', LIBDIR)
+  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
 
 
@@ -36,18 +58,35 @@ if Config::CONFIG['target_os'] == 'mingw32'
   unless find_header('libxslt/libxslt.h', header)
     abort "need libxslt"
   end
+  unless find_header('libexslt/libexslt.h', header)
+    abort "need libexslt"
+  end
 
   header = File.join(ROOT, 'cross', 'iconv-1.9.2.win32', 'include')
   unless find_header('iconv.h', header)
     abort "need iconv"
   end
 else
-  unless find_header('libxml/xmlversion.h',
-                     File.join(INCLUDEDIR, "libxml2"), '/usr/include/libxml2'
-                    )
+  HEADER_DIRS = [
+    File.join(INCLUDEDIR, "libxml2"),
+    INCLUDEDIR,
+    '/usr/include/libxml2',
+    '/usr/local/include/libxml2'
+  ]
+
+  [
+    '/opt/local/include/libxml2',
+    '/opt/local/include',
+  ].each { |x| HEADER_DIRS.unshift(x) } if use_macports
+
+  unless find_header('libxml/xmlversion.h', *HEADER_DIRS)
     abort "need libxml"
   end
-  unless find_header('libxslt/xslt.h', INCLUDEDIR, '/usr/include')
+
+  unless find_header('libxslt/xslt.h', *HEADER_DIRS)
+    abort "need libxslt"
+  end
+  unless find_header('libexslt/exslt.h', *HEADER_DIRS)
     abort "need libxslt"
   end
 

data/ext/nokogiri/html_sax_parser.c

@@ -3,22 +3,26 @@
 static VALUE native_parse_file(VALUE self, VALUE data, VALUE encoding)
 {
   xmlSAXHandlerPtr handler;
+  htmlDocPtr hdoc ;
   Data_Get_Struct(self, xmlSAXHandler, handler);
-  htmlSAXParseFile( StringValuePtr(data),
-                    (const char *)StringValuePtr(encoding),
-                    (htmlSAXHandlerPtr)handler,
-                    (void *)self );
+  hdoc = htmlSAXParseFile( StringValuePtr(data),
+                           (const char *)StringValuePtr(encoding),
+                           (htmlSAXHandlerPtr)handler,
+                           (void *)self );
+  xmlFreeDoc(hdoc);
   return data;
 }
 
 static VALUE native_parse_memory(VALUE self, VALUE data, VALUE encoding)
 {
   xmlSAXHandlerPtr handler;
+  htmlDocPtr hdoc ;
   Data_Get_Struct(self, xmlSAXHandler, handler);
-  htmlSAXParseDoc(  (xmlChar *)StringValuePtr(data),
-                    (const char *)StringValuePtr(encoding),
-                    (htmlSAXHandlerPtr)handler,
-                    (void *)self );
+  hdoc = htmlSAXParseDoc(  (xmlChar *)StringValuePtr(data),
+                           (const char *)StringValuePtr(encoding),
+                           (htmlSAXHandlerPtr)handler,
+                           (void *)self );
+  xmlFreeDoc(hdoc);
   return data;
 }
 

data/ext/nokogiri/xml_node.c

@@ -70,14 +70,20 @@ static VALUE internal_subset(VALUE self)
  * call-seq:
  *  dup
  *
- * Copy this node
+ * Copy this node.  An optional depth may be passed in, but it defaults
+ * to a deep copy.  0 is a shallow copy, 1 is a deep copy.
  */
-static VALUE duplicate_node(VALUE self)
+static VALUE duplicate_node(int argc, VALUE *argv, VALUE self)
 {
+  VALUE level;
+
+  if(rb_scan_args(argc, argv, "01", &level) == 0)
+    level = INT2NUM(1);
+
   xmlNodePtr node, dup;
   Data_Get_Struct(self, xmlNode, node);
 
-  dup = xmlCopyNode(node, 1);
+  dup = xmlCopyNode(node, NUM2INT(level));
   if(dup == NULL) return Qnil;
   dup->doc = node->doc;
   assert(node->parent);
@@ -150,12 +156,7 @@ static VALUE previous_sibling(VALUE self)
   return Nokogiri_wrap_xml_node(sibling);
 }
 
-/*
- *  call-seq:
- *    replace(new_node)
- *
- *  replace node with the new node in the document.
- */
+/* :nodoc: */
 static VALUE replace(VALUE self, VALUE _new_node)
 {
   xmlNodePtr node, new_node;
@@ -338,18 +339,20 @@ static VALUE get_content(VALUE self)
 
 /*
  * call-seq:
- *  parent=(parent_node)
+ *  add_child(node)
  *
- * Set the parent Node for this Node
+ * Add +node+ as a child of this node. Returns the new child node.
  */
-static VALUE set_parent(VALUE self, VALUE parent_node)
+static VALUE add_child(VALUE self, VALUE child)
 {
-  xmlNodePtr node, parent;
-  Data_Get_Struct(self, xmlNode, node);
-  Data_Get_Struct(parent_node, xmlNode, parent);
+  xmlNodePtr node, parent, new_child;
+  Data_Get_Struct(child, xmlNode, node);
+  Data_Get_Struct(self, xmlNode, parent);
 
-  xmlAddChild(parent, node);
-  return parent_node;
+  if(!(new_child = xmlAddChild(parent, node)))
+    rb_raise(rb_eRuntimeError, "Could not add new child");
+
+  return Nokogiri_wrap_xml_node(new_child);
 }
 
 /*
@@ -443,13 +446,20 @@ static VALUE add_next_sibling(VALUE self, VALUE rb_node)
 static VALUE add_previous_sibling(VALUE self, VALUE rb_node)
 {
   xmlNodePtr node, new_sibling;
+  Check_Type(rb_node, T_DATA);
+
   Data_Get_Struct(self, xmlNode, node);
   Data_Get_Struct(rb_node, xmlNode, new_sibling);
-  xmlAddPrevSibling(node, new_sibling);
+
+  if(!(new_sibling = xmlAddPrevSibling(node, new_sibling)))
+    rb_raise(rb_eRuntimeError, "Could not add previous sibling");
 
   rb_funcall(rb_node, rb_intern("decorate!"), 0);
 
-  return rb_node;
+  VALUE rb_new_sibling = Nokogiri_wrap_xml_node(new_sibling);
+  rb_funcall(rb_new_sibling, rb_intern("decorate!"), 0);
+
+  return rb_new_sibling;
 }
 
 /*
@@ -458,16 +468,21 @@ static VALUE add_previous_sibling(VALUE self, VALUE rb_node)
  *
  * Returns this node as XML
  */
-static VALUE to_xml(VALUE self)
+static VALUE to_xml(int argc, VALUE *argv, VALUE self)
 {
   xmlBufferPtr buf ;
   xmlNodePtr node ;
-  VALUE xml ;
+  VALUE xml, level;
+
+  if(rb_scan_args(argc, argv, "01", &level) == 0)
+    level = INT2NUM(1);
+
+  Check_Type(level, T_FIXNUM);
 
   Data_Get_Struct(self, xmlNode, node);
 
   buf = xmlBufferCreate() ;
-  xmlNodeDump(buf, node->doc, node, 2, 1);
+  xmlNodeDump(buf, node->doc, node, 2, NUM2INT(level));
   xml = rb_str_new2((char*)buf->content);
   xmlBufferFree(buf);
   return xml ;
@@ -646,12 +661,11 @@ void init_xml_node()
 
   rb_define_method(klass, "name", get_name, 0);
   rb_define_method(klass, "name=", set_name, 1);
-  rb_define_method(klass, "parent=", set_parent, 1);
+  rb_define_method(klass, "add_child", add_child, 1);
   rb_define_method(klass, "parent", get_parent, 0);
   rb_define_method(klass, "child", child, 0);
   rb_define_method(klass, "next_sibling", next_sibling, 0);
   rb_define_method(klass, "previous_sibling", previous_sibling, 0);
-  rb_define_method(klass, "replace", replace, 1);
   rb_define_method(klass, "type", type, 0);
   rb_define_method(klass, "content", get_content, 0);
   rb_define_method(klass, "path", path, 0);
@@ -664,12 +678,13 @@ void init_xml_node()
   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, 0);
-  rb_define_method(klass, "dup", duplicate_node, 0);
+  rb_define_method(klass, "to_xml", to_xml, -1);
+  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);
   rb_define_method(klass, "pointer_id", pointer_id, 0);
 
+  rb_define_private_method(klass, "replace_with_node", replace, 1);
   rb_define_private_method(klass, "native_content=", set_content, 1);
   rb_define_private_method(klass, "get", get, 1);
 }

data/ext/nokogiri/xml_node_set.c

@@ -28,6 +28,9 @@ static VALUE push(VALUE self, VALUE rb_node)
   xmlNodeSetPtr node_set;
   xmlNodePtr node;
 
+  if(! rb_funcall(rb_node, rb_intern("is_a?"), 1, cNokogiriXmlNode))
+    rb_raise(rb_eArgError, "node must be a Nokogiri::XML::Node");
+
   Data_Get_Struct(self, xmlNodeSet, node_set);
   Data_Get_Struct(rb_node, xmlNode, node);
   xmlXPathNodeSetAdd(node_set, node);

data/ext/nokogiri/xml_sax_parser.c

@@ -42,6 +42,7 @@ static VALUE native_parse_io(VALUE self, VALUE io, VALUE encoding)
       enc
   );
   xmlParseDocument(sax_ctx);
+  xmlFreeParserCtxt(sax_ctx);
   return io;
 }
 

data/ext/nokogiri/xml_xpath_context.c

@@ -25,18 +25,132 @@ static VALUE register_ns(VALUE self, VALUE prefix, VALUE uri)
   return self;
 }
 
+static void ruby_funcall(xmlXPathParserContextPtr ctx, int nargs)
+{
+  VALUE xpath_handler = Qnil;
+  xmlXPathObjectPtr obj;
+
+  assert(ctx);
+  assert(ctx->context);
+  assert(ctx->context->userData);
+  assert(ctx->context->doc);
+  assert(ctx->context->doc->_private);
+
+  xpath_handler = (VALUE)(ctx->context->userData);
+
+  VALUE * argv = (VALUE *)calloc((unsigned int)nargs, sizeof(VALUE));
+  VALUE doc = (VALUE)ctx->context->doc->_private;
+
+  int i = nargs - 1;
+  do {
+    obj = valuePop(ctx);
+    switch(obj->type) {
+      case XPATH_STRING:
+        argv[i] = rb_str_new2((char *)obj->stringval);
+        break;
+      case XPATH_BOOLEAN:
+        argv[i] = obj->boolval == 1 ? Qtrue : Qfalse;
+        break;
+      case XPATH_NUMBER:
+        argv[i] = rb_float_new(obj->floatval);
+        break;
+      case XPATH_NODESET:
+        argv[i] = Nokogiri_wrap_xml_node_set(obj->nodesetval);
+        break;
+      default:
+        argv[i] = rb_str_new2((char *)xmlXPathCastToString(obj));
+    }
+    xmlXPathFreeNodeSetList(obj);
+  } while(i-- > 0);
+
+  VALUE result = rb_funcall2(
+      xpath_handler,
+      rb_intern((const char *)ctx->context->function),
+      nargs,
+      argv
+  );
+  free(argv);
+
+  VALUE node_set = Qnil;
+  xmlNodeSetPtr xml_node_set = NULL;
+
+  switch(TYPE(result)) {
+    case T_FLOAT:
+    case T_BIGNUM:
+    case T_FIXNUM:
+      xmlXPathReturnNumber(ctx, NUM2DBL(result));
+      break;
+    case T_STRING:
+      xmlXPathReturnString(
+          ctx,
+          xmlXPathWrapCString(StringValuePtr(result))
+      );
+      break;
+    case T_TRUE:
+      xmlXPathReturnTrue(ctx);
+      break;
+    case T_FALSE:
+      xmlXPathReturnFalse(ctx);
+      break;
+    case T_NIL:
+      break;
+    case T_ARRAY:
+      node_set = rb_funcall(
+          cNokogiriXmlNodeSet,
+          rb_intern("new"),
+          2,
+          doc,
+          result
+      );
+      Data_Get_Struct(node_set, xmlNodeSet, xml_node_set);
+      xmlXPathReturnNodeSet(ctx, xmlXPathNodeSetMerge(NULL, xml_node_set));
+      break;
+    case T_DATA:
+      if(rb_funcall(result, rb_intern("is_a?"), 1, cNokogiriXmlNodeSet)) {
+        Data_Get_Struct(result, xmlNodeSet, xml_node_set);
+        // Copy the node set, otherwise it will get GC'd.
+        xmlXPathReturnNodeSet(ctx, xmlXPathNodeSetMerge(NULL, xml_node_set));
+        break;
+      }
+    default:
+      rb_raise(rb_eRuntimeError, "Invalid return type");
+  }
+}
+
+static xmlXPathFunction lookup( void *ctx,
+                                const xmlChar * name,
+                                const xmlChar* ns_uri )
+{
+  VALUE xpath_handler = (VALUE)ctx;
+  if(rb_respond_to(xpath_handler, rb_intern((const char *)name)))
+    return ruby_funcall;
+
+  return NULL;
+}
+
 /*
  * call-seq:
  *  evaluate(search_path)
  *
  * Evaluate the +search_path+ returning an XML::XPath object.
  */
-static VALUE evaluate(VALUE self, VALUE search_path)
+static VALUE evaluate(int argc, VALUE *argv, VALUE self)
 {
+  VALUE search_path, xpath_handler;
   xmlXPathContextPtr ctx;
   Data_Get_Struct(self, xmlXPathContext, ctx);
 
+  if(rb_scan_args(argc, argv, "11", &search_path, &xpath_handler) == 1)
+    xpath_handler = Qnil;
+
   xmlChar* query = (xmlChar *)StringValuePtr(search_path);
+
+  if(Qnil != xpath_handler) {
+    // FIXME: not sure if this is the correct place to shove private data.
+    ctx->userData = (void *)xpath_handler;
+    xmlXPathRegisterFuncLookup(ctx, lookup, (void *)xpath_handler);
+  }
+
   xmlXPathObjectPtr xpath = xmlXPathEvalExpression(query, ctx);
   if(xpath == NULL) {
     VALUE xpath = rb_const_get(mNokogiriXml, rb_intern("XPath"));
@@ -72,8 +186,10 @@ static VALUE new(VALUE klass, VALUE nodeobj)
   Data_Get_Struct(nodeobj, xmlNode, node);
 
   xmlXPathContextPtr ctx = xmlXPathNewContext(node->doc);
-  ctx->node = node ;
-  return Data_Wrap_Struct(klass, 0, deallocate, ctx);
+  ctx->node = node;
+  VALUE self = Data_Wrap_Struct(klass, 0, deallocate, ctx);
+  //rb_iv_set(self, "@xpath_handler", Qnil);
+  return self;
 }
 
 VALUE cNokogiriXmlXpathContext;
@@ -94,6 +210,6 @@ void init_xml_xpath_context(void)
   cNokogiriXmlXpathContext = klass;
 
   rb_define_singleton_method(klass, "new", new, 1);
-  rb_define_method(klass, "evaluate", evaluate, 1);
+  rb_define_method(klass, "evaluate", evaluate, -1);
   rb_define_method(klass, "register_ns", register_ns, 2);
 }