nokogiri 1.4.7 → 1.5.0.beta.1

data/bin/nokogiri

@@ -7,7 +7,6 @@ require 'rubygems'
 require 'nokogiri'
 
 parse_class = Nokogiri
-encoding = nil
 
 opts = OptionParser.new do |opts|
   opts.banner = "Nokogiri: an HTML, XML, SAX, and Reader parser"
@@ -23,10 +22,6 @@ opts = OptionParser.new do |opts|
     parse_class = {:xml => Nokogiri::XML, :html => Nokogiri::HTML}[v]
   end
 
-  opts.on("-E", "--encoding encoding", "Read as encoding (default #{encoding})") do |v|
-    encoding = v
-  end
-
   opts.on_tail("-?", "--help", "Show this message") do
     puts opts
     exit
@@ -47,7 +42,7 @@ if uri.to_s.strip.empty?
   exit 1
 end
 
-@doc = parse_class.parse(open(uri).read, nil, encoding)
+@doc = parse_class.parse(open(uri).read)
 
 puts "Your document is stored in @doc..."
 IRB.start

data/deps.rip

@@ -0,0 +1,5 @@
+git://github.com/luislavena/rake-compiler.git v0.6.0
+git://github.com/tenderlove/rexical.git
+git://github.com/jimweirich/rake.git
+git://github.com/jmhodges/fakehoe.git
+git://github.com/tenderlove/racc.git

data/ext/nokogiri/extconf.rb

@@ -17,7 +17,7 @@ end
 $CFLAGS << " #{ENV["CFLAGS"]}"
 $LIBS << " #{ENV["LIBS"]}"
 
-if Config::CONFIG['target_os'] == 'mingw32' || Config::CONFIG['target_os'] =~ /mswin32/
+if Config::CONFIG['target_os'] == 'mingw32'
   $CFLAGS << " -DXP_WIN -DXP_WIN32 -DUSE_INCLUDED_VASPRINTF"
 elsif Config::CONFIG['target_os'] =~ /solaris/
   $CFLAGS << " -DUSE_INCLUDED_VASPRINTF"
@@ -29,79 +29,58 @@ if RbConfig::MAKEFILE_CONFIG['CC'] =~ /mingw/
   $CFLAGS << " -DIN_LIBXML"
 end
 
-if RbConfig::MAKEFILE_CONFIG['CC'] =~ /gcc/
-  $CFLAGS << " -O3 -Wall -Wcast-qual -Wwrite-strings -Wconversion -Wmissing-noreturn -Winline"
-end
-
-if Config::CONFIG['target_os'] =~ /mswin32/
-  lib_prefix = 'lib'
-
-  # There's no default include/lib dir on Windows. Let's just add the Ruby ones
-  # and resort on the search path specified by INCLUDE and LIB environment
-  # variables
-  HEADER_DIRS = [INCLUDEDIR]
-  LIB_DIRS = [LIBDIR]
-  XML2_HEADER_DIRS = [File.join(INCLUDEDIR, "libxml2"), INCLUDEDIR]
-
-else
-  lib_prefix = ''
-
-  HEADER_DIRS = [
-    # First search /opt/local for macports
-    '/opt/local/include',
+$CFLAGS << " -O3 -Wall -Wcast-qual -Wwrite-strings -Wconversion -Wmissing-noreturn -Winline"
 
-    # Then search /usr/local for people that installed from source
-    '/usr/local/include',
+HEADER_DIRS = [
+  # First search /opt/local for macports
+  '/opt/local/include',
 
-    # Check the ruby install locations
-    INCLUDEDIR,
+  # Then search /usr/local for people that installed from source
+  '/usr/local/include',
 
-    # Finally fall back to /usr
-    '/usr/include',
-    '/usr/include/libxml2',
-  ]
+  # Check the ruby install locations
+  INCLUDEDIR,
 
-  LIB_DIRS = [
-    # First search /opt/local for macports
-    '/opt/local/lib',
+  # Finally fall back to /usr
+  '/usr/include',
+  '/usr/include/libxml2',
+]
 
-    # Then search /usr/local for people that installed from source
-    '/usr/local/lib',
+LIB_DIRS = [
+  # First search /opt/local for macports
+  '/opt/local/lib',
 
-    # Check the ruby install locations
-    LIBDIR,
+  # Then search /usr/local for people that installed from source
+  '/usr/local/lib',
 
-    # Finally fall back to /usr
-    '/usr/lib',
-  ]
+  # Check the ruby install locations
+  LIBDIR,
 
-  XML2_HEADER_DIRS = [
-    '/opt/local/include/libxml2',
-    '/usr/local/include/libxml2',
-    File.join(INCLUDEDIR, "libxml2")
-  ] + HEADER_DIRS
-end
+  # Finally fall back to /usr
+  '/usr/lib',
+]
 
 dir_config('zlib', HEADER_DIRS, LIB_DIRS)
 dir_config('iconv', HEADER_DIRS, LIB_DIRS)
-dir_config('xml2', XML2_HEADER_DIRS, LIB_DIRS)
+dir_config('xml2', [
+           '/opt/local/include/libxml2',
+           '/usr/local/include/libxml2',
+           File.join(INCLUDEDIR, "libxml2")] + HEADER_DIRS, LIB_DIRS)
 dir_config('xslt', HEADER_DIRS, LIB_DIRS)
 
 def asplode(lib)
   abort "-----\n#{lib} is missing.  please visit http://nokogiri.org/tutorials/installing_nokogiri.html for help with installing dependencies.\n-----"
 end
 
-pkg_config('libxslt') if RUBY_PLATFORM =~ /mingw/
-
+asplode "iconv"    unless find_header('iconv.h')
 asplode "libxml2"  unless find_header('libxml/parser.h')
 asplode "libxslt"  unless find_header('libxslt/xslt.h')
 asplode "libexslt" unless find_header('libexslt/exslt.h')
-asplode "libiconv" unless have_func('iconv_open', 'iconv.h') or have_library('iconv', 'iconv_open', 'iconv.h')
-asplode "libxml2"  unless find_library("#{lib_prefix}xml2", 'xmlParseDoc')
-asplode "libxslt"  unless find_library("#{lib_prefix}xslt", 'xsltParseStylesheetDoc')
-asplode "libexslt" unless find_library("#{lib_prefix}exslt", 'exsltFuncRegister')
+asplode "zlib"     unless find_library('z', 'gzopen')
+asplode "libxml2"  unless find_library('xml2', 'xmlParseDoc')
+asplode "libxslt"  unless find_library('xslt', 'xsltParseStylesheetDoc')
+asplode "libexslt" unless find_library('exslt', 'exsltFuncRegister')
 
-have_func 'xmlFirstElementChild'
 have_func('xmlRelaxNGSetParserStructuredErrors')
 have_func('xmlRelaxNGSetParserStructuredErrors')
 have_func('xmlRelaxNGSetValidStructuredErrors')

data/ext/nokogiri/nokogiri.c

@@ -30,14 +30,12 @@ int is_2_6_16(void)
 
 void Init_nokogiri()
 {
-#ifndef __MACRUBY__
   xmlMemSetup(
       (xmlFreeFunc)ruby_xfree,
       (xmlMallocFunc)ruby_xmalloc,
       (xmlReallocFunc)ruby_xrealloc,
       strdup
   );
-#endif
 
   mNokogiri         = rb_define_module("Nokogiri");
   mNokogiriXml      = rb_define_module_under(mNokogiri, "XML");

data/ext/nokogiri/nokogiri.h

@@ -27,7 +27,6 @@ int vasprintf (char **strp, const char *fmt, va_list ap);
 #include <libxml/HTMLparser.h>
 #include <libxml/HTMLtree.h>
 #include <libxml/relaxng.h>
-#include <libxslt/extensions.h>
 #include <ruby.h>
 
 #ifdef HAVE_RUBY_ENCODING_H
@@ -78,8 +77,6 @@ int is_2_6_16(void) ;
 #define RBSTR_OR_QNIL(_str) \
   (_str ? NOKOGIRI_STR_NEW2(_str) : Qnil)
 
-#include <xml_libxml2_hacks.h>
-
 #include <xml_io.h>
 #include <xml_document.h>
 #include <html_entity_lookup.h>
@@ -135,8 +132,6 @@ extern VALUE mNokogiriXslt ;
 #define NOKOGIRI_DEBUG_START(p)
 #define NOKOGIRI_DEBUG_END(p)
 
-#endif
-
 #ifndef RSTRING_PTR
 #define RSTRING_PTR(s) (RSTRING(s)->ptr)
 #endif
@@ -153,10 +148,6 @@ extern VALUE mNokogiriXslt ;
 #define RARRAY_LEN(a) RARRAY(a)->len
 #endif
 
-#ifndef __builtin_expect
-# if defined(__GNUC__)
-#  define __builtin_expect(expr, c) __builtin_expect((long)(expr), (long)(c))
-# endif
 #endif
 
 #endif

data/ext/nokogiri/xml_document.c

@@ -41,25 +41,11 @@ static void dealloc(xmlDocPtr doc)
 static void recursively_remove_namespaces_from_node(xmlNodePtr node)
 {
   xmlNodePtr child ;
-  xmlAttrPtr property ;
 
   xmlSetNs(node, NULL);
 
   for (child = node->children ; child ; child = child->next)
     recursively_remove_namespaces_from_node(child);
-
-  if (node->nsDef) {
-    xmlFreeNsList(node->nsDef);
-    node->nsDef = NULL;
-  }
-
-  if (node->properties != NULL) {
-    property = node->properties ;
-    while (property != NULL) {
-      if (property->ns) property->ns = NULL ;
-      property = property->next ;
-    }
-  }
 }
 
 /*

data/ext/nokogiri/xml_dtd.c

@@ -84,9 +84,9 @@ static VALUE attributes(VALUE self)
 
   Data_Get_Struct(self, xmlDtd, dtd);
 
-  hash = rb_hash_new();
+  if(!dtd->attributes) return Qnil;
 
-  if(!dtd->attributes) return hash;
+  hash = rb_hash_new();
 
   xmlHashScan((xmlHashTablePtr)dtd->attributes, element_copier, (void *)hash);
 

data/ext/nokogiri/xml_io.c

@@ -2,47 +2,22 @@
 
 static ID id_read, id_write;
 
-VALUE read_check(VALUE *args) {
-  return rb_funcall(args[0], id_read, 1, args[1]);
-}
-
-VALUE read_failed(void) {
-	return Qnil;
-}
-
 int io_read_callback(void * ctx, char * buffer, int len) {
-  VALUE string, args[2];
-  size_t str_len, safe_len;
-
-  args[0] = (VALUE)ctx;
-  args[1] = INT2NUM(len);
-
-  string = rb_rescue(read_check, (VALUE)args, read_failed, 0);
+  VALUE io = (VALUE)ctx;
+  VALUE string = rb_funcall(io, id_read, 1, INT2NUM(len));
 
   if(NIL_P(string)) return 0;
 
-  str_len = (size_t)RSTRING_LEN(string);
-  safe_len = str_len > (size_t)len ? (size_t)len : str_len;
-  memcpy(buffer, StringValuePtr(string), safe_len);
+  memcpy(buffer, StringValuePtr(string), (unsigned int)RSTRING_LEN(string));
 
-  return safe_len;
-}
-
-VALUE write_check(VALUE *args) {
-  return rb_funcall(args[0], id_write, 1, args[1]);
-}
-
-VALUE write_failed(void) {
-	return Qnil;
+  return (int)RSTRING_LEN(string);
 }
 
 int io_write_callback(void * ctx, char * buffer, int len) {
-  VALUE args[2];
-
-  args[0] = (VALUE)ctx;
-  args[1] = rb_str_new(buffer, (long)len);
+  VALUE io = (VALUE)ctx;
+  VALUE string = rb_str_new(buffer, len);
 
-  rb_rescue(write_check, (VALUE)args, write_failed, 0);
+  rb_funcall(io, id_write, 1, string);
   return len;
 }
 

data/ext/nokogiri/xml_node.c

@@ -14,7 +14,10 @@ static void debug_node_dealloc(xmlNodePtr x)
 
 static void mark(xmlNodePtr node)
 {
-  rb_gc_mark(DOC_RUBY_OBJECT(node->doc));
+  /* it's OK if the document isn't fully realized (as in XML::Reader). */
+  /* see http://github.com/tenderlove/nokogiri/issues/closed/#issue/95 */
+  if (DOC_RUBY_OBJECT_TEST(node->doc) && DOC_RUBY_OBJECT(node->doc))
+    rb_gc_mark(DOC_RUBY_OBJECT(node->doc));
 }
 
 /* :nodoc: */
@@ -100,7 +103,7 @@ 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 ;
 
   if(!rb_obj_is_kind_of(reparentee_obj, cNokogiriXmlNode))
     rb_raise(rb_eArgError, "node must be a Nokogiri::XML::Node");
@@ -138,34 +141,6 @@ static VALUE reparent_node_with(VALUE pivot_obj, VALUE reparentee_obj, pivot_rep
     }
   }
 
-  if (reparentee->type == XML_TEXT_NODE && pivot->next && pivot->next->type == XML_TEXT_NODE) {
-    /*
-     *  libxml merges text nodes in a right-to-left fashion, meaning that if
-     *  there are two text nodes who would be adjacent, the right (or following,
-     *  or next) node will be merged into the left (or preceding, or previous)
-     *  node.
-     *
-     *  and by "merged" I mean the string contents will be concatenated onto the
-     *  left node's contents, and then the node will be freed.
-     *
-     *  which means that if we have a ruby object wrapped around the right node,
-     *  its memory would be freed out from under it.
-     *
-     *  so, we detect this edge case and unlink-and-root the text node before it gets
-     *  merged. then we dup the node and insert that duplicate back into the
-     *  document where the real node was.
-     *
-     *  yes, this is totally lame.
-     */
-    next_text     = pivot->next ;
-    new_next_text = xmlDocCopyNode(next_text, pivot->doc, 1) ;
-
-    xmlUnlinkNode(next_text);
-    NOKOGIRI_ROOT_NODE(next_text);
-
-    xmlAddNextSibling(pivot, new_next_text);
-  }
-
   /* TODO: I really want to remove this.  We shouldn't support 2.6.16 anymore */
   if ( reparentee->type == XML_TEXT_NODE && pivot->type == XML_TEXT_NODE && is_2_6_16() ) {
     /* work around a string-handling bug in libxml 2.6.16. we'd rather leak than segfault. */
@@ -248,13 +223,7 @@ static VALUE encode_special_chars(VALUE self, VALUE string)
  * call-seq:
  *  create_internal_subset(name, external_id, system_id)
  *
- * Create the internal subset of a document.
- *
- *   doc.create_internal_subset("chapter", "-//OASIS//DTD DocBook XML//EN", "chapter.dtd")
- *   # => <!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML//EN" "chapter.dtd">
- *
- *   doc.create_internal_subset("chapter", nil, "chapter.dtd")
- *   # => <!DOCTYPE chapter SYSTEM "chapter.dtd">
+ * Create an internal subset
  */
 static VALUE create_internal_subset(VALUE self, VALUE name, VALUE external_id, VALUE system_id)
 {
@@ -380,8 +349,6 @@ static VALUE duplicate_node(int argc, VALUE *argv, VALUE self)
   dup = xmlDocCopyNode(node, node->doc, (int)NUM2INT(level));
   if(dup == NULL) return Qnil;
 
-  NOKOGIRI_ROOT_NODE(dup);
-
   return Nokogiri_wrap_xml_node(rb_obj_class(self), dup);
 }
 
@@ -779,8 +746,7 @@ 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 Nokogiri::XML::Namespace for the node, if one exists.
  */
 static VALUE namespace(VALUE self)
 {
@@ -797,7 +763,7 @@ static VALUE namespace(VALUE self)
  *  call-seq:
  *    namespace_definitions()
  *
- *  returns namespaces defined on self element directly, as an array of Namespace objects. Includes both a default namespace (as in"xmlns="), and prefixed namespaces (as in "xmlns:prefix=").
+ *  returns a list of Namespace nodes defined on _self_
  */
 static VALUE namespace_definitions(VALUE self)
 {
@@ -826,10 +792,8 @@ static VALUE namespace_definitions(VALUE self)
  *  call-seq:
  *    namespace_scopes()
  *
- * returns namespaces in scope for self -- those defined on self element
- * directly or any ancestor node -- as an array of Namespace objects.  Default
- * namespaces ("xmlns=" style) for self are included in this array; Default
- * namespaces for  ancestors, however, are not. See also #namespaces
+ *  returns a list of Namespace nodes in scope for _self_. this is all
+ *  namespaces defined in the node, or in any ancestor node.
  */
 static VALUE namespace_scopes(VALUE self)
 {
@@ -1053,41 +1017,30 @@ static VALUE line(VALUE self)
  * call-seq:
  *  add_namespace_definition(prefix, href)
  *
- * Adds a namespace definition with +prefix+ using +href+ value. The result is
- * as if parsed XML for this node had included an attribute
- * 'xmlns:prefix=value'.  A default namespace for this node ("xmlns=") can be
- * added by passing 'nil' for prefix. Namespaces added this way will not
- * show up in #attributes, but they will be included as an xmlns attribute
- * when the node is serialized to XML.
+ * Adds a namespace definition with +prefix+ using +href+
  */
 static VALUE add_namespace_definition(VALUE self, VALUE prefix, VALUE href)
 {
-  xmlNodePtr node, namespacee;
+  xmlNodePtr node;
   xmlNsPtr ns;
 
   Data_Get_Struct(self, xmlNode, node);
-  namespacee = node ;
 
-  ns = xmlSearchNs(
-      node->doc,
+  ns = xmlNewNs(
       node,
+      (const xmlChar *)StringValuePtr(href),
       (const xmlChar *)(NIL_P(prefix) ? NULL : StringValuePtr(prefix))
   );
 
   if(!ns) {
-    if (node->type != XML_ELEMENT_NODE) {
-      namespacee = node->parent;
-    }
-    ns = xmlNewNs(
-        namespacee,
-        (const xmlChar *)StringValuePtr(href),
+    ns = xmlSearchNs(
+        node->doc,
+        node,
         (const xmlChar *)(NIL_P(prefix) ? NULL : StringValuePtr(prefix))
     );
   }
 
-  if (!ns) return Qnil ;
-
-  if(NIL_P(prefix) || node != namespacee) xmlSetNs(node, ns);
+  if(NIL_P(prefix)) xmlSetNs(node, ns);
 
   return Nokogiri_wrap_xml_namespace(node->doc, ns);
 }
@@ -1168,13 +1121,14 @@ static VALUE in_context(VALUE self, VALUE _str, VALUE _options)
 {
   xmlNodePtr node;
   xmlNodePtr list;
-  xmlNodePtr child_iter;
   xmlNodeSetPtr set;
-  xmlParserErrors error;
   VALUE doc, err;
 
   Data_Get_Struct(self, xmlNode, node);
 
+  if(!node->parent)
+    rb_raise(rb_eRuntimeError, "no contextual parsing on unlinked nodes");
+
   doc = DOC_RUBY_OBJECT(node->doc);
   err = rb_iv_get(doc, "@errors");
 
@@ -1187,41 +1141,19 @@ static VALUE in_context(VALUE self, VALUE _str, VALUE _options)
   htmlHandleOmittedElem(0);
 #endif
 
-  error = xmlParseInNodeContext(
+  xmlParseInNodeContext(
       node,
       StringValuePtr(_str),
       (int)RSTRING_LEN(_str),
       (int)NUM2INT(_options),
       &list);
 
-  /* make sure parent/child pointers are coherent so an unlink will work properly (#331) */
-  child_iter = node->doc->children ;
-  while (child_iter) {
-    if (child_iter->parent != (xmlNodePtr)node->doc)
-      child_iter->parent = (xmlNodePtr)node->doc ;
-    child_iter = child_iter->next ;
-  }
-
 #ifndef HTML_PARSE_NOIMPLIED
   htmlHandleOmittedElem(1);
 #endif
 
   xmlSetStructuredErrorFunc(NULL, NULL);
 
-  /* FIXME: This probably needs to handle more constants... */
-  switch(error) {
-    case XML_ERR_OK:
-      break;
-
-    case XML_ERR_INTERNAL_ERROR:
-    case XML_ERR_NO_MEMORY:
-      rb_raise(rb_eRuntimeError, "error parsing fragment (%d)", error);
-      break;
-
-    default:
-      break;
-  }
-
   set = xmlXPathNodeSetCreate(NULL);
 
   while(list) {
@@ -1238,8 +1170,6 @@ VALUE Nokogiri_wrap_xml_node(VALUE klass, xmlNodePtr node)
   VALUE document = Qnil ;
   VALUE node_cache = Qnil ;
   VALUE rb_node = Qnil ;
-  int node_has_a_document = 0 ;
-  void (*mark_method)(xmlNodePtr) = NULL ;
 
   assert(node);
 
@@ -1248,9 +1178,11 @@ VALUE Nokogiri_wrap_xml_node(VALUE klass, xmlNodePtr node)
 
   if(NULL != node->_private) return (VALUE)node->_private;
 
-  if(!RTEST(klass)) {
-    switch(node->type)
-    {
+  if(RTEST(klass))
+    rb_node = Data_Wrap_Struct(klass, mark, debug_node_dealloc, node) ;
+
+  else switch(node->type)
+  {
     case XML_ELEMENT_NODE:
       klass = cNokogiriXmlElement;
       break;
@@ -1289,19 +1221,15 @@ VALUE Nokogiri_wrap_xml_node(VALUE klass, xmlNodePtr node)
       break;
     default:
       klass = cNokogiriXmlNode;
-    }
   }
 
-  /* It's OK if the node doesn't have a fully-realized document (as in XML::Reader). */
-  /* see https://github.com/tenderlove/nokogiri/issues/95 */
-  /* and https://github.com/tenderlove/nokogiri/issues/439 */
-  node_has_a_document = (DOC_RUBY_OBJECT_TEST(node->doc) && DOC_RUBY_OBJECT(node->doc)) ? 1 : 0 ;
-  mark_method = node_has_a_document ? mark : NULL ;
+  rb_node = Data_Wrap_Struct(klass, mark, debug_node_dealloc, node) ;
 
-  rb_node = Data_Wrap_Struct(klass, mark_method, debug_node_dealloc, node) ;
   node->_private = (void *)rb_node;
 
-  if (node_has_a_document) {
+  if (DOC_RUBY_OBJECT_TEST(node->doc) && DOC_RUBY_OBJECT(node->doc)) {
+    /* it's OK if the document isn't fully realized (as in XML::Reader). */
+    /* see http://github.com/tenderlove/nokogiri/issues/closed/#issue/95 */
     document = DOC_RUBY_OBJECT(node->doc);
     node_cache = DOC_NODE_CACHE(node->doc);
     rb_ary_push(node_cache, rb_node);