nokogiri 1.6.1-java → 1.6.2-java

data/ext/nokogiri/html_document.c

@@ -77,7 +77,7 @@ static VALUE read_io( VALUE klass,
 
     error = xmlGetLastError();
     if(error)
-      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
     else
       rb_raise(rb_eRuntimeError, "Could not parse document");
 
@@ -123,7 +123,7 @@ static VALUE read_memory( VALUE klass,
 
     error = xmlGetLastError();
     if(error)
-      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
     else
       rb_raise(rb_eRuntimeError, "Could not parse document");
 

data/ext/nokogiri/nokogiri.c

@@ -14,7 +14,12 @@ VALUE mNokogiriHtmlSax ;
  */
 int vasprintf (char **strp, const char *fmt, va_list ap)
 {
-  int len = vsnprintf (NULL, 0, fmt, ap) + 1;
+  /* Mingw32/64 have a broken vsnprintf implementation that fails when
+   * using a zero-byte limit in order to retrieve the required size for malloc.
+   * So we use a one byte buffer instead.
+   */
+  char tmp[1];
+  int len = vsnprintf (tmp, 1, fmt, ap) + 1;
   char *res = (char *)malloc((unsigned int)len);
   if (res == NULL)
       return -1;

data/ext/nokogiri/xml_document.c

@@ -231,7 +231,7 @@ static VALUE read_io( VALUE klass,
 
     error = xmlGetLastError();
     if(error)
-      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
     else
       rb_raise(rb_eRuntimeError, "Could not parse document");
 
@@ -275,7 +275,7 @@ static VALUE read_memory( VALUE klass,
 
     error = xmlGetLastError();
     if(error)
-      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
     else
       rb_raise(rb_eRuntimeError, "Could not parse document");
 
@@ -417,7 +417,7 @@ static VALUE create_entity(int argc, VALUE *argv, VALUE self)
   if(NULL == ptr) {
     xmlErrorPtr error = xmlGetLastError();
     if(error)
-      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
     else
       rb_raise(rb_eRuntimeError, "Could not create entity");
 
@@ -497,6 +497,7 @@ static VALUE canonicalize(int argc, VALUE* argv, VALUE self)
     ns = NULL;
   }
   else{
+    Check_Type(incl_ns, T_ARRAY);
     ns_len = RARRAY_LEN(incl_ns);
     ns = calloc((size_t)ns_len+1, sizeof(xmlChar *));
     for (i = 0 ; i < ns_len ; i++) {
@@ -510,7 +511,7 @@ static VALUE canonicalize(int argc, VALUE* argv, VALUE self)
   xmlC14NExecute(doc, cb, ctx, 
     (int)      (NIL_P(mode)        ? 0 : NUM2INT(mode)), 
     ns,
-    (int)      (NIL_P(with_comments)        ? 0 : 1),
+    (int)      RTEST(with_comments),
     buf);
 
   xmlOutputBufferClose(buf);

data/ext/nokogiri/xml_node.c

@@ -14,7 +14,14 @@ static void debug_node_dealloc(xmlNodePtr x)
 
 static void mark(xmlNodePtr node)
 {
-  rb_gc_mark(DOC_RUBY_OBJECT(node->doc));
+  xmlNodePtr doc = node->doc;
+  if(doc->type == XML_DOCUMENT_NODE || doc->type == XML_HTML_DOCUMENT_NODE) {
+    if(DOC_RUBY_OBJECT_TEST(doc)) {
+      rb_gc_mark(DOC_RUBY_OBJECT(doc));
+    }
+  } else if(node->doc->_private) {
+    rb_gc_mark((VALUE)doc->_private);
+  }
 }
 
 /* :nodoc: */
@@ -132,7 +139,7 @@ static xmlNodePtr xmlReplaceNodeWrapper(xmlNodePtr pivot, xmlNodePtr new_node)
 static VALUE reparent_node_with(VALUE pivot_obj, VALUE reparentee_obj, pivot_reparentee_func prf)
 {
   VALUE reparented_obj ;
-  xmlNodePtr reparentee, pivot, reparented, next_text, new_next_text ;
+  xmlNodePtr reparentee, pivot, reparented, next_text, new_next_text, parent ;
 
   if(!rb_obj_is_kind_of(reparentee_obj, cNokogiriXmlNode))
     rb_raise(rb_eArgError, "node must be a Nokogiri::XML::Node");
@@ -142,9 +149,71 @@ static VALUE reparent_node_with(VALUE pivot_obj, VALUE reparentee_obj, pivot_rep
   Data_Get_Struct(reparentee_obj, xmlNode, reparentee);
   Data_Get_Struct(pivot_obj, xmlNode, pivot);
 
-  if(XML_DOCUMENT_NODE == reparentee->type || XML_HTML_DOCUMENT_NODE == reparentee->type)
-    rb_raise(rb_eArgError, "cannot reparent a document node");
+  /*
+   * Check if nodes given are appropriate to have a parent-child
+   * relationship, based on the DOM specification.
+   *
+   * cf. http://www.w3.org/TR/2004/REC-DOM-Level-3-Core-20040407/core.html#ID-1590626202
+   */
+  if (prf == xmlAddChild) {
+    parent = pivot;
+  } else {
+    parent = pivot->parent;
+  }
+
+  if (parent) {
+    switch (parent->type) {
+    case XML_DOCUMENT_NODE:
+    case XML_HTML_DOCUMENT_NODE:
+      switch (reparentee->type) {
+        case XML_ELEMENT_NODE:
+        case XML_PI_NODE:
+        case XML_COMMENT_NODE:
+        case XML_DOCUMENT_TYPE_NODE:
+      /*
+       * The DOM specification says no to adding text-like nodes
+       * directly to a document, but we allow it for compatibility.
+       */
+        case XML_TEXT_NODE:
+        case XML_CDATA_SECTION_NODE:
+        case XML_ENTITY_REF_NODE:
+          goto ok;
+      }
+      break;
+    case XML_DOCUMENT_FRAG_NODE:
+    case XML_ENTITY_REF_NODE:
+    case XML_ELEMENT_NODE:
+      switch (reparentee->type) {
+      case XML_ELEMENT_NODE:
+      case XML_PI_NODE:
+      case XML_COMMENT_NODE:
+      case XML_TEXT_NODE:
+      case XML_CDATA_SECTION_NODE:
+      case XML_ENTITY_REF_NODE:
+        goto ok;
+      }
+      break;
+    case XML_ATTRIBUTE_NODE:
+      switch (reparentee->type) {
+      case XML_TEXT_NODE:
+      case XML_ENTITY_REF_NODE:
+        goto ok;
+      }
+      break;
+    case XML_TEXT_NODE:
+      /*
+       * xmlAddChild() breaks the DOM specification in that it allows
+       * adding a text node to another, in which case text nodes are
+       * coalesced, but since our JRuby version does not support such
+       * operation, we should inhibit it.
+       */
+      break;
+    }
+
+    rb_raise(rb_eArgError, "cannot reparent %s there", rb_obj_classname(reparentee_obj));
+  }
 
+ok:
   xmlUnlinkNode(reparentee);
 
   if (reparentee->doc != pivot->doc || reparentee->type == XML_TEXT_NODE) {
@@ -850,8 +919,8 @@ static VALUE attribute_nodes(VALUE self)
  *  call-seq:
  *    namespace()
  *
- *  returns the default namespace set on this node (as with an "xmlns="
- *  attribute), as a Namespace object.
+ *  returns the namespace of the element or attribute node as a Namespace
+ *  object, or nil if there is no namespace for the element or attribute.
  */
 static VALUE namespace(VALUE self)
 {
@@ -1258,7 +1327,7 @@ static VALUE process_xincludes(VALUE self, VALUE options)
 
     error = xmlGetLastError();
     if(error)
-      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
     else
       rb_raise(rb_eRuntimeError, "Could not perform xinclude substitution");
   }

data/ext/nokogiri/xml_reader.c

@@ -471,7 +471,7 @@ static VALUE read_more(VALUE self)
 
   error = xmlGetLastError();
   if(error)
-    rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+    rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
   else
     rb_raise(rb_eRuntimeError, "Error pulling: %d", ret);
 

data/ext/nokogiri/xml_sax_parser_context.c

@@ -201,6 +201,44 @@ static VALUE column(VALUE self)
   return Qnil;
 }
 
+/*
+ * call-seq:
+ *  recovery=(boolean)
+ *
+ * Should this parser recover from structural errors? It will not stop processing
+ * file on structural errors if if set to true
+ */
+static VALUE set_recovery(VALUE self, VALUE value)
+{
+  xmlParserCtxtPtr ctxt;
+  Data_Get_Struct(self, xmlParserCtxt, ctxt);
+
+  if(value == Qfalse)
+    ctxt->recovery = 0;
+  else
+    ctxt->recovery = 1;
+
+  return value;
+}
+
+/*
+ * call-seq:
+ *  recovery
+ *
+ * Should this parser recover from structural errors? It will not stop processing
+ * file on structural errors if if set to true
+ */
+static VALUE get_recovery(VALUE self)
+{
+  xmlParserCtxtPtr ctxt;
+  Data_Get_Struct(self, xmlParserCtxt, ctxt);
+
+  if(ctxt->recovery == 0)
+    return Qfalse;
+  else
+    return Qtrue;
+}
+
 void init_xml_sax_parser_context()
 {
   VALUE nokogiri  = rb_define_module("Nokogiri");
@@ -217,6 +255,8 @@ void init_xml_sax_parser_context()
   rb_define_method(klass, "parse_with", parse_with, 1);
   rb_define_method(klass, "replace_entities=", set_replace_entities, 1);
   rb_define_method(klass, "replace_entities", get_replace_entities, 0);
+  rb_define_method(klass, "recovery=", set_recovery, 1);
+  rb_define_method(klass, "recovery", get_recovery, 0);
   rb_define_method(klass, "line", line, 0);
   rb_define_method(klass, "column", column, 0);
 }

data/ext/nokogiri/xml_syntax_error.c

@@ -3,19 +3,24 @@
 void Nokogiri_error_array_pusher(void * ctx, xmlErrorPtr error)
 {
   VALUE list = (VALUE)ctx;
-  rb_ary_push(list,  Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+  rb_ary_push(list,  Nokogiri_wrap_xml_syntax_error(error));
 }
 
 void Nokogiri_error_raise(void * ctx, xmlErrorPtr error)
 {
-  rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+  rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
 }
 
-VALUE Nokogiri_wrap_xml_syntax_error(VALUE klass, xmlErrorPtr error)
+VALUE Nokogiri_wrap_xml_syntax_error(xmlErrorPtr error)
 {
-  VALUE msg, e;
+  VALUE msg, e, klass;
 
-  if(!klass) klass = cNokogiriXmlSyntaxError;
+  klass = cNokogiriXmlSyntaxError;
+
+  if (error->domain == XML_FROM_XPATH) {
+    VALUE xpath = rb_const_get(mNokogiriXml, rb_intern("XPath"));
+    klass = rb_const_get(xpath, rb_intern("SyntaxError"));
+  }
 
   msg = (error && error->message) ? NOKOGIRI_STR_NEW2(error->message) : Qnil;
 

data/ext/nokogiri/xml_syntax_error.h

@@ -4,7 +4,7 @@
 #include <nokogiri.h>
 
 void init_xml_syntax_error();
-VALUE Nokogiri_wrap_xml_syntax_error(VALUE klass, xmlErrorPtr error);
+VALUE Nokogiri_wrap_xml_syntax_error(xmlErrorPtr error);
 void Nokogiri_error_array_pusher(void * ctx, xmlErrorPtr error);
 NORETURN(void Nokogiri_error_raise(void * ctx, xmlErrorPtr error));
 

data/ext/nokogiri/xml_xpath_context.c

@@ -171,15 +171,6 @@ static xmlXPathFunction lookup( void *ctx,
   return NULL;
 }
 
-NORETURN(static void xpath_exception_handler(void * ctx, xmlErrorPtr error));
-static void xpath_exception_handler(void * ctx, xmlErrorPtr error)
-{
-  VALUE xpath = rb_const_get(mNokogiriXml, rb_intern("XPath"));
-  VALUE klass = rb_const_get(xpath, rb_intern("SyntaxError"));
-
-  rb_exc_raise(Nokogiri_wrap_xml_syntax_error(klass, error));
-}
-
 NORETURN(static void xpath_generic_exception_handler(void * ctx, const char *msg, ...));
 static void xpath_generic_exception_handler(void * ctx, const char *msg, ...)
 {
@@ -221,7 +212,7 @@ static VALUE evaluate(int argc, VALUE *argv, VALUE self)
   }
 
   xmlResetLastError();
-  xmlSetStructuredErrorFunc(NULL, xpath_exception_handler);
+  xmlSetStructuredErrorFunc(NULL, Nokogiri_error_raise);
 
   /* For some reason, xmlXPathEvalExpression will blow up with a generic error */
   /* when there is a non existent function. */
@@ -232,11 +223,8 @@ static VALUE evaluate(int argc, VALUE *argv, VALUE self)
   xmlSetGenericErrorFunc(NULL, NULL);
 
   if(xpath == NULL) {
-    VALUE xpath = rb_const_get(mNokogiriXml, rb_intern("XPath"));
-    VALUE klass = rb_const_get(xpath, rb_intern("SyntaxError"));
-
     xmlErrorPtr error = xmlGetLastError();
-    rb_exc_raise(Nokogiri_wrap_xml_syntax_error(klass, error));
+    rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));
   }
 
   assert(ctx->doc);

data/ext/nokogiri/xslt_stylesheet.c

@@ -237,7 +237,7 @@ static void shutdownFunc(xsltTransformContextPtr ctxt,
  *  call-seq:
  *    register(uri, custom_handler_class)
  *
- *  Register a class that implements custom XLST transformation functions.
+ *  Register a class that implements custom XSLT transformation functions.
  */
 static VALUE registr(VALUE self, VALUE uri, VALUE obj)
 {

data/lib/nokogiri.rb

@@ -2,9 +2,6 @@
 # Modify the PATH on windows so that the external DLLs will get loaded.
 
 require 'rbconfig'
-ENV['PATH'] = [File.expand_path(
-  File.join(File.dirname(__FILE__), "..", "ext", "nokogiri")
-), ENV['PATH']].compact.join(';') if RbConfig::CONFIG['host_os'] =~ /(mswin|mingw)/i
 
 if defined?(RUBY_ENGINE) && RUBY_ENGINE == "jruby"
   # The line below caused a problem on non-GAE rack environment.
@@ -13,7 +10,7 @@ if defined?(RUBY_ENGINE) && RUBY_ENGINE == "jruby"
   # However, simply cutting defined?(JRuby::Rack::VERSION) off resulted in
   # an unable-to-load-nokogiri problem. Thus, now, Nokogiri checks the presense
   # of appengine-rack.jar in $LOAD_PATH. If Nokogiri is on GAE, Nokogiri
-  # should skip loading xml jars. This is because those are in WEB-INF/lib and 
+  # should skip loading xml jars. This is because those are in WEB-INF/lib and
   # already set in the classpath.
   unless $LOAD_PATH.to_s.include?("appengine-rack")
     require 'stringio'
@@ -25,7 +22,12 @@ if defined?(RUBY_ENGINE) && RUBY_ENGINE == "jruby"
   end
 end
 
-require 'nokogiri/nokogiri'
+begin
+  RUBY_VERSION =~ /(\d+.\d+)/
+  require "nokogiri/#{$1}/nokogiri"
+rescue LoadError
+  require 'nokogiri/nokogiri'
+end
 require 'nokogiri/version'
 require 'nokogiri/syntax_error'
 require 'nokogiri/xml'
@@ -36,7 +38,8 @@ require 'nokogiri/css'
 require 'nokogiri/html/builder'
 
 # Nokogiri parses and searches XML/HTML very quickly, and also has
-# correctly implemented CSS3 selector support as well as XPath support.
+# correctly implemented CSS3 selector support as well as XPath 1.0
+# support.
 #
 # Parsing a document returns either a Nokogiri::XML::Document, or a
 # Nokogiri::HTML::Document depending on the kind of document you parse.
@@ -65,20 +68,19 @@ module Nokogiri
     ###
     # Parse an HTML or XML document.  +string+ contains the document.
     def parse string, url = nil, encoding = nil, options = nil
-      doc =
-        if string.respond_to?(:read) ||
-          string =~ /^\s*<[^Hh>]*html/i # Probably html
-          Nokogiri.HTML(
-            string,
-            url,
-            encoding, options || XML::ParseOptions::DEFAULT_HTML
-          )
-        else
-          Nokogiri.XML(string, url, encoding,
-                        options || XML::ParseOptions::DEFAULT_XML)
-        end
-      yield doc if block_given?
-      doc
+      if string.respond_to?(:read) ||
+          /^\s*<(?:!DOCTYPE\s+)?html[\s>]/i === string[0, 512]
+        # Expect an HTML indicator to appear within the first 512
+        # characters of a document. (<?xml ?> + <?xml-stylesheet ?>
+        # shouldn't be that long)
+        Nokogiri.HTML(string, url, encoding,
+          options || XML::ParseOptions::DEFAULT_HTML)
+      else
+        Nokogiri.XML(string, url, encoding,
+          options || XML::ParseOptions::DEFAULT_XML)
+      end.tap { |doc|
+        yield doc if block_given?
+      }
     end
 
     ###
@@ -110,6 +112,14 @@ module Nokogiri
       Nokogiri(*args, &block).slop!
     end
   end
+
+  # Make sure to support some popular encoding aliases not known by
+  # all iconv implementations.
+  {
+    'Windows-31J' => 'CP932',	# Windows-31J is the IANA registered name of CP932.
+  }.each { |alias_name, name|
+    EncodingHandler.alias(name, alias_name) if EncodingHandler[alias_name].nil?
+  }
 end
 
 ###
@@ -120,8 +130,7 @@ end
 # To specify the type of document, use Nokogiri.XML or Nokogiri.HTML.
 def Nokogiri(*args, &block)
   if block_given?
-    builder = Nokogiri::HTML::Builder.new(&block)
-    return builder.doc.root
+    Nokogiri::HTML::Builder.new(&block).doc.root
   else
     Nokogiri.parse(*args)
   end

data/lib/nokogiri/css/node.rb

@@ -22,60 +22,10 @@ module Nokogiri
       ###
       # Convert this CSS node to xpath with +prefix+ using +visitor+
       def to_xpath prefix = '//', visitor = XPathVisitor.new
-        self.preprocess!
         prefix = '.' if ALLOW_COMBINATOR_ON_SELF.include?(type) && value.first.nil?
         prefix + visitor.accept(self)
       end
 
-      # Preprocess this node tree
-      def preprocess!
-        ### Deal with nth-child
-        matches = find_by_type(
-          [:CONDITIONAL_SELECTOR,
-            [:ELEMENT_NAME],
-            [:PSEUDO_CLASS,
-              [:FUNCTION]
-            ]
-          ]
-        )
-        matches.each do |match|
-          if match.value[1].value[0].value[0] =~ /^nth-(last-)?child/
-            tag_name = match.value[0].value.first
-            match.value[0].value = ['*']
-            match.value[1] = Node.new(:COMBINATOR, [
-              match.value[1].value[0],
-              Node.new(:FUNCTION, ['self(', tag_name])
-            ])
-          end
-        end
-
-        ### Deal with first-child, last-child
-        matches = find_by_type(
-          [:CONDITIONAL_SELECTOR,
-            [:ELEMENT_NAME], [:PSEUDO_CLASS]
-        ])
-        matches.each do |match|
-          if ['first-child', 'last-child'].include?(match.value[1].value.first)
-            which = match.value[1].value.first.gsub(/-\w*$/, '')
-            tag_name = match.value[0].value.first
-            match.value[0].value = ['*']
-            match.value[1] = Node.new(:COMBINATOR, [
-              Node.new(:FUNCTION, ["#{which}("]),
-              Node.new(:FUNCTION, ['self(', tag_name])
-            ])
-          elsif 'only-child' == match.value[1].value.first
-            tag_name = match.value[0].value.first
-            match.value[0].value = ['*']
-            match.value[1] = Node.new(:COMBINATOR, [
-              Node.new(:FUNCTION, ["#{match.value[1].value.first}("]),
-              Node.new(:FUNCTION, ['self(', tag_name])
-            ])
-          end
-        end
-
-        self
-      end
-
       # Find a node by type using +types+
       def find_by_type types
         matches = []