nokogiri 1.5.5.rc3-java → 1.5.6-java

data/ext/java/nokogiri/internals/XmlDomParserContext.java

@@ -72,6 +72,8 @@ public class XmlDomParserContext extends ParserContext {
         "http://apache.org/xml/features/nonvalidating/load-dtd-grammar";
     protected static final String FEATURE_INCLUDE_IGNORABLE_WHITESPACE =
         "http://apache.org/xml/features/dom/include-ignorable-whitespace";
+    protected static final String CONTINUE_AFTER_FATAL_ERROR =
+        "http://apache.org/xml/features/continue-after-fatal-error";
     protected static final String FEATURE_NOT_EXPAND_ENTITY =
         "http://apache.org/xml/features/dom/create-entity-ref-nodes";
     protected static final String FEATURE_VALIDATION = "http://xml.org/sax/features/validation";
@@ -110,17 +112,21 @@ public class XmlDomParserContext extends ParserContext {
             System.setProperty("org.apache.xerces.xni.parser.XMLParserConfiguration",
                     "org.apache.xerces.parsers.XIncludeParserConfiguration");
         }
-        
-        parser = new XmlDomParser(options);
+
+        parser = new NokogiriDomParser(options);
         parser.setErrorHandler(errorHandler);
 
         // Fix for Issue#586.  This limits entity expansion up to 100000 and nodes up to 3000.
-        setProperty(SECURITY_MANAGER, new org.apache.xerces.util.SecurityManager());        
-        
+        setProperty(SECURITY_MANAGER, new org.apache.xerces.util.SecurityManager());
+
         if (options.noBlanks) {
             setFeature(FEATURE_INCLUDE_IGNORABLE_WHITESPACE, false);
         }
 
+        if (options.recover) {
+            setFeature(CONTINUE_AFTER_FATAL_ERROR, true);
+        }
+
         if (options.dtdValid) {
             setFeature(FEATURE_VALIDATION, true);
         }

data/ext/nokogiri/extconf.rb

@@ -27,6 +27,7 @@ end
 
 if RbConfig::MAKEFILE_CONFIG['CC'] =~ /mingw/
   $CFLAGS << " -DIN_LIBXML"
+  $LIBS << " -lz" # TODO why is this necessary?
 end
 
 if RbConfig::MAKEFILE_CONFIG['CC'] =~ /gcc/

data/ext/nokogiri/xml_document.c

@@ -156,7 +156,7 @@ static VALUE set_encoding(VALUE self, VALUE encoding)
   Data_Get_Struct(self, xmlDoc, doc);
 
   if (doc->encoding)
-      free((char *) doc->encoding); // this may produce a gcc cast warning
+      free((char *) doc->encoding); /* this may produce a gcc cast warning */
 
   doc->encoding = xmlStrdup((xmlChar *)StringValuePtr(encoding));
 
@@ -367,7 +367,7 @@ static VALUE new(int argc, VALUE *argv, VALUE klass)
  *
  *  For more information on why this probably is *not* a good thing in general,
  *  please direct your browser to
- *  http://tenderlovemaking.com/2009/04/23/namespaces-in-xml/
+ *  http://tenderlovemaking.com/2009/04/23/namespaces-in-xml.html
  */
 VALUE remove_namespaces_bang(VALUE self)
 {

data/ext/nokogiri/xml_node.c

@@ -84,7 +84,7 @@ static xmlNodePtr xmlReplaceNodeWrapper(xmlNodePtr pivot, xmlNodePtr new_node)
   }
 
   /* work around libxml2 issue: https://bugzilla.gnome.org/show_bug.cgi?id=615612 */
-  if (retval->type == XML_TEXT_NODE) {
+  if (retval && retval->type == XML_TEXT_NODE) {
     if (retval->prev && retval->prev->type == XML_TEXT_NODE) {
       retval = xmlTextMerge(retval->prev, retval);
     }
@@ -699,23 +699,40 @@ static VALUE set(VALUE self, VALUE property, VALUE value)
  *
  * Get the value for +attribute+
  */
-static VALUE get(VALUE self, VALUE attribute)
+static VALUE get(VALUE self, VALUE rattribute)
 {
   xmlNodePtr node;
-  xmlChar* propstr ;
-  VALUE rval ;
-  Data_Get_Struct(self, xmlNode, node);
+  xmlChar* value = 0;
+  VALUE rvalue ;
+  char* attribute = 0;
+  char *colon = 0, *attr_name = 0, *prefix = 0;
+  xmlNsPtr ns;
 
-  if(NIL_P(attribute)) return Qnil;
+  if (NIL_P(rattribute)) return Qnil;
 
-  propstr = xmlGetProp(node, (xmlChar *)StringValuePtr(attribute));
+  Data_Get_Struct(self, xmlNode, node);
+  attribute = strdup(StringValuePtr(rattribute));
+
+  colon = strchr(attribute, ':');
+  if (colon) {
+    (*colon) = 0 ; /* create two null-terminated strings of the prefix and attribute name */
+    prefix = attribute ;
+    attr_name = colon + 1 ;
+    ns = xmlSearchNs(node->doc, node, (const xmlChar *)(prefix));
+    if (ns) {
+      value = xmlGetNsProp(node, (xmlChar*)(attr_name), ns->href);
+    }
+  } else {
+    value = xmlGetNoNsProp(node, (xmlChar*)attribute);
+  }
 
-  if(!propstr) return Qnil;
+  free(attribute);
+  if (!value) return Qnil;
 
-  rval = NOKOGIRI_STR_NEW2(propstr);
+  rvalue = NOKOGIRI_STR_NEW2(value);
+  xmlFree(value);
 
-  xmlFree(propstr);
-  return rval ;
+  return rvalue ;
 }
 
 /*
@@ -892,7 +909,7 @@ static VALUE node_type(VALUE self)
  *
  * Set the content for this Node
  */
-static VALUE set_content(VALUE self, VALUE content)
+static VALUE native_content(VALUE self, VALUE content)
 {
   xmlNodePtr node, child, next ;
   Data_Get_Struct(self, xmlNode, node);
@@ -1288,7 +1305,7 @@ static VALUE in_context(VALUE self, VALUE _str, VALUE _options)
       child_iter = node;
       while (child_iter->parent)
         child_iter = child_iter->parent;
-	
+
       if (child_iter->type == XML_DOCUMENT_FRAG_NODE)
         node->doc->children = NULL;
     }
@@ -1458,6 +1475,7 @@ void init_xml_node()
   rb_define_method(klass, "create_external_subset", create_external_subset, 3);
   rb_define_method(klass, "pointer_id", pointer_id, 0);
   rb_define_method(klass, "line", line, 0);
+  rb_define_method(klass, "native_content=", native_content, 1);
 
   rb_define_private_method(klass, "process_xincludes", process_xincludes, 1);
   rb_define_private_method(klass, "in_context", in_context, 2);
@@ -1467,7 +1485,6 @@ void init_xml_node()
   rb_define_private_method(klass, "replace_node", replace, 1);
   rb_define_private_method(klass, "dump_html", dump_html, 0);
   rb_define_private_method(klass, "native_write_to", native_write_to, 4);
-  rb_define_private_method(klass, "native_content=", set_content, 1);
   rb_define_private_method(klass, "get", get, 1);
   rb_define_private_method(klass, "set", set, 2);
   rb_define_private_method(klass, "set_namespace", set_namespace, 1);

data/ext/nokogiri/xml_sax_parser.c

@@ -7,6 +7,7 @@ static ID id_start_document, id_end_document, id_start_element, id_end_element;
 static ID id_start_element_namespace, id_end_element_namespace;
 static ID id_comment, id_characters, id_xmldecl, id_error, id_warning;
 static ID id_cdata_block, id_cAttribute;
+static ID id_processing_instruction;
 
 #define STRING_OR_NULL(str) \
    (RTEST(str) ? StringValuePtr(str) : NULL)
@@ -236,6 +237,19 @@ static void cdata_block(void * ctx, const xmlChar * value, int len)
   rb_funcall(doc, id_cdata_block, 1, string);
 }
 
+static void processing_instruction(void * ctx, const xmlChar * name, const xmlChar * content)
+{
+  VALUE self = NOKOGIRI_SAX_SELF(ctx);
+  VALUE doc = rb_iv_get(self, "@document");
+
+  rb_funcall( doc,
+              id_processing_instruction,
+              2,
+              NOKOGIRI_STR_NEW2(name),
+              NOKOGIRI_STR_NEW2(content)
+  );
+}
+
 static void deallocate(xmlSAXHandlerPtr handler)
 {
   NOKOGIRI_DEBUG_START(handler);
@@ -260,6 +274,7 @@ static VALUE allocate(VALUE klass)
   handler->warning = warning_func;
   handler->error = error_func;
   handler->cdataBlock = cdata_block;
+  handler->processingInstruction = processing_instruction;
   handler->initialized = XML_SAX2_MAGIC;
 
   return Data_Wrap_Struct(klass, NULL, deallocate, handler);
@@ -290,4 +305,5 @@ void init_xml_sax_parser()
   id_cAttribute     = rb_intern("Attribute");
   id_start_element_namespace = rb_intern("start_element_namespace");
   id_end_element_namespace = rb_intern("end_element_namespace");
+  id_processing_instruction = rb_intern("processing_instruction");
 }

data/ext/nokogiri/xslt_stylesheet.c

@@ -109,6 +109,10 @@ static VALUE serialize(VALUE self, VALUE xmlobj)
     return rval ;
 }
 
+static void swallow_superfluous_xml_errors(void * userdata, xmlErrorPtr error, ...)
+{
+}
+
 /*
  *  call-seq:
  *    transform(document, params = [])
@@ -126,12 +130,13 @@ static VALUE serialize(VALUE self, VALUE xmlobj)
  */
 static VALUE transform(int argc, VALUE* argv, VALUE self)
 {
-    VALUE xmldoc, paramobj ;
+    VALUE xmldoc, paramobj, errstr, exception ;
     xmlDocPtr xml ;
     xmlDocPtr result ;
     nokogiriXsltStylesheetTuple *wrapper;
     const char** params ;
     long param_len, j ;
+    int parse_error_occurred ;
 
     rb_scan_args(argc, argv, "11", &xmldoc, &paramobj);
     if (NIL_P(paramobj)) { paramobj = rb_ary_new2(0L) ; }
@@ -158,10 +163,22 @@ static VALUE transform(int argc, VALUE* argv, VALUE self)
     }
     params[param_len] = 0 ;
 
+    errstr = rb_str_new(0, 0);
+    xsltSetGenericErrorFunc((void *)errstr, xslt_generic_error_handler);
+    xmlSetGenericErrorFunc(NULL, (xmlGenericErrorFunc)&swallow_superfluous_xml_errors);
+
     result = xsltApplyStylesheet(wrapper->ss, xml, params);
     free(params);
 
-    if (!result) rb_raise(rb_eRuntimeError, "could not perform xslt transform on document");
+    xsltSetGenericErrorFunc(NULL, NULL);
+    xmlSetGenericErrorFunc(NULL, NULL);
+
+    parse_error_occurred = (Qfalse == rb_funcall(errstr, rb_intern("empty?"), 0));
+
+    if (parse_error_occurred) {
+      exception = rb_exc_new3(rb_eRuntimeError, errstr);
+      rb_exc_raise(exception);
+    }
 
     return Nokogiri_wrap_xml_document((VALUE)0, result) ;
 }

data/lib/nokogiri/version.rb

@@ -1,6 +1,6 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.5.5.rc3'
+  VERSION = '1.5.6'
 
   class VersionInfo # :nodoc:
     def jruby?
@@ -52,6 +52,9 @@ module Nokogiri
         hash_info['libxml']['compiled']  = compiled_parser_version
         hash_info['libxml']['loaded']    = loaded_parser_version
         hash_info['warnings']            = warnings
+      elsif jruby?
+        hash_info['xerces']   = Nokogiri::XERCES_VERSION
+        hash_info['nekohtml'] = Nokogiri::NEKO_VERSION
       end
 
       hash_info

data/lib/nokogiri/xml/document.rb

@@ -149,13 +149,15 @@ module Nokogiri
       # Non-prefixed default namespaces (as in "xmlns=") are not included
       # in the hash.
       #
-      # Note this is a very expensive operation in current implementation, as it
-      # traverses the entire graph, and also has to bring each node across the
-      # libxml bridge into a ruby object.
+      # Note that this method does an xpath lookup for nodes with
+      # namespaces, and as a result the order may be dependent on the
+      # implementation of the underlying XML library.
+      #
       def collect_namespaces
-        ns = {}
-        traverse { |j| ns.merge!(j.namespaces) }
-        ns
+        xpath("//namespace::*").inject({}) do |hash, ns|
+          hash[["xmlns",ns.prefix].compact.join(":")] = ns.href if ns.prefix != "xml"
+          hash
+        end
       end
 
       # Get the list of decorators given +key+

data/lib/nokogiri/xml/document_fragment.rb

@@ -13,7 +13,8 @@ module Nokogiri
         children = if ctx
                      # Fix for issue#490
                      if Nokogiri.jruby?
-                       ctx.parse("<root>#{tags}</root>").xpath("/root/node()")
+                       # fix for issue #770
+                       ctx.parse("<root #{namespace_declarations(ctx)}>#{tags}</root>").children
                      else
                        ctx.parse(tags)
                      end
@@ -93,6 +94,14 @@ module Nokogiri
 
       private
 
+      # fix for issue 770
+      def namespace_declarations ctx
+        ctx.namespace_scopes.map do |namespace|
+          prefix = namespace.prefix.nil? ? "" : ":#{namespace.prefix}"
+          %Q{xmlns#{prefix}="#{namespace.href}"}
+        end.join ' '
+      end
+
       def coerce data
         return super unless String === data
 

data/lib/nokogiri/xml/node.rb

@@ -251,14 +251,13 @@ module Nokogiri
       ###
       # Get the attribute value for the attribute +name+
       def [] name
-        return nil unless key?(name.to_s)
         get(name.to_s)
       end
 
       ###
       # Set the attribute value for the attribute +name+ to +value+
       def []= name, value
-        set name.to_s, value
+        set name.to_s, value.to_s
       end
 
       ###
@@ -299,20 +298,7 @@ module Nokogiri
       def add_previous_sibling node_or_tags
         raise ArgumentError.new("A document may not have multiple root nodes.") if parent.is_a?(XML::Document) && !node_or_tags.is_a?(XML::ProcessingInstruction)
 
-        node_or_tags = coerce(node_or_tags)
-        if node_or_tags.is_a?(XML::NodeSet)
-          if text?
-            pivot = Nokogiri::XML::Node.new 'dummy', document
-            add_previous_sibling_node pivot
-          else
-            pivot = self
-          end
-          node_or_tags.each { |n| pivot.send :add_previous_sibling_node, n }
-          pivot.unlink if text?
-        else
-          add_previous_sibling_node node_or_tags
-        end
-        node_or_tags
+        add_sibling :previous, node_or_tags
       end
 
       ###
@@ -325,20 +311,7 @@ module Nokogiri
       def add_next_sibling node_or_tags
         raise ArgumentError.new("A document may not have multiple root nodes.") if parent.is_a?(XML::Document)
         
-        node_or_tags = coerce(node_or_tags)
-        if node_or_tags.is_a?(XML::NodeSet)
-          if text?
-            pivot = Nokogiri::XML::Node.new 'dummy', document
-            add_next_sibling_node pivot
-          else
-            pivot = self
-          end
-          node_or_tags.reverse_each { |n| pivot.send :add_next_sibling_node, n }
-          pivot.unlink if text?
-        else
-          add_next_sibling_node node_or_tags
-        end
-        node_or_tags
+        add_sibling :next, node_or_tags
       end
 
       ####
@@ -403,17 +376,22 @@ module Nokogiri
       #
       # Also see related method +swap+.
       def replace node_or_tags
+        # We cannot replace a text node directly, otherwise libxml will return
+        # an internal error at parser.c:13031, I don't know exactly why
+        # libxml is trying to find a parent node that is an element or document
+        # so I can't tell if this is bug in libxml or not. issue #775.
+        if text?
+          replacee = Nokogiri::XML::Node.new 'dummy', document
+          add_previous_sibling_node replacee
+          unlink
+          return replacee.replace node_or_tags
+        end
+
         node_or_tags = coerce(node_or_tags)
+
         if node_or_tags.is_a?(XML::NodeSet)
-          if text?
-            replacee = Nokogiri::XML::Node.new 'dummy', document
-            add_previous_sibling_node replacee
-            unlink
-          else
-            replacee = self
-          end
-          node_or_tags.each { |n| replacee.add_previous_sibling n }
-          replacee.unlink
+          node_or_tags.each { |n| add_previous_sibling n }
+          unlink
         else
           replace_node node_or_tags
         end
@@ -761,12 +739,7 @@ module Nokogiri
       # See Node#write_to for a list of +options+.  For formatted output,
       # use Node#to_xhtml instead.
       def to_html options = {}
-        # FIXME: this is a hack around broken libxml versions
-        return dump_html if Nokogiri.uses_libxml? && %w[2 6] === LIBXML_VERSION.split('.')[0..1]
-
-        options[:save_with] |= SaveOptions::DEFAULT_HTML if options[:save_with]
-        options[:save_with] = SaveOptions::DEFAULT_HTML unless options[:save_with]
-        serialize(options)
+        to_format SaveOptions::DEFAULT_HTML, options
       end
 
       ###
@@ -787,12 +760,7 @@ module Nokogiri
       #
       # See Node#write_to for a list of +options+
       def to_xhtml options = {}
-        # FIXME: this is a hack around broken libxml versions
-        return dump_html if Nokogiri.uses_libxml? && %w[2 6] === LIBXML_VERSION.split('.')[0..1]
-
-        options[:save_with] |= SaveOptions::DEFAULT_XHTML if options[:save_with]
-        options[:save_with] = SaveOptions::DEFAULT_XHTML unless options[:save_with]
-        serialize(options)
+        to_format SaveOptions::DEFAULT_XHTML, options
       end
 
       ###
@@ -835,11 +803,7 @@ module Nokogiri
       #
       # See Node#write_to for a list of +options+
       def write_html_to io, options = {}
-        # FIXME: this is a hack around broken libxml versions
-        return (io << dump_html) if Nokogiri.uses_libxml? && %w[2 6] === LIBXML_VERSION.split('.')[0..1]
-
-        options[:save_with] ||= SaveOptions::DEFAULT_HTML
-        write_to io, options
+        write_format_to SaveOptions::DEFAULT_HTML, io, options
       end
 
       ###
@@ -847,11 +811,7 @@ module Nokogiri
       #
       # See Node#write_to for a list of +options+
       def write_xhtml_to io, options = {}
-        # FIXME: this is a hack around broken libxml versions
-        return (io << dump_html) if Nokogiri.uses_libxml? && %w[2 6] === LIBXML_VERSION.split('.')[0..1]
-
-        options[:save_with] ||= SaveOptions::DEFAULT_XHTML
-        write_to io, options
+        write_format_to SaveOptions::DEFAULT_XHTML, io, options
       end
 
       ###
@@ -898,6 +858,43 @@ module Nokogiri
 
       private
 
+      def add_sibling next_or_previous, node_or_tags
+        impl = (next_or_previous == :next) ? :add_next_sibling_node : :add_previous_sibling_node
+        iter = (next_or_previous == :next) ? :reverse_each          : :each
+
+        node_or_tags = coerce node_or_tags
+        if node_or_tags.is_a?(XML::NodeSet)
+          if text?
+            pivot = Nokogiri::XML::Node.new 'dummy', document
+            send impl, pivot
+          else
+            pivot = self
+          end
+          node_or_tags.send(iter) { |n| pivot.send impl, n }
+          pivot.unlink if text?
+        else
+          send impl, node_or_tags
+        end
+        node_or_tags
+      end
+
+      def to_format save_option, options
+        # FIXME: this is a hack around broken libxml versions
+        return dump_html if Nokogiri.uses_libxml? && %w[2 6] === LIBXML_VERSION.split('.')[0..1]
+
+        options[:save_with] |= save_option if options[:save_with]
+        options[:save_with] = save_option unless options[:save_with]
+        serialize(options)
+      end
+
+      def write_format_to save_option, io, options
+        # FIXME: this is a hack around broken libxml versions
+        return (io << dump_html) if Nokogiri.uses_libxml? && %w[2 6] === LIBXML_VERSION.split('.')[0..1]
+
+        options[:save_with] ||= save_option
+        write_to io, options
+      end
+
       def extract_params params # :nodoc:
         # Pop off our custom function handler if it exists
         handler = params.find { |param|