rubyjedi-nokogiri_java 1.4.0.20100513161003-java

data/ext/nokogiri/xml_attr.h

@@ -0,0 +1,9 @@
+#ifndef NOKOGIRI_XML_ATTR
+#define NOKOGIRI_XML_ATTR
+
+#include <nokogiri.h>
+
+void init_xml_attr();
+
+extern VALUE cNokogiriXmlAttr;
+#endif

data/ext/nokogiri/xml_attribute_decl.c

@@ -0,0 +1,67 @@
+#include <xml_attribute_decl.h>
+
+/*
+ * call-seq:
+ *  attribute_type
+ *
+ * The attribute_type for this AttributeDecl
+ */
+static VALUE attribute_type(VALUE self)
+{
+  xmlAttributePtr node;
+  Data_Get_Struct(self, xmlAttribute, node);
+  return INT2NUM((long)node->atype);
+}
+
+/*
+ * call-seq:
+ *  default
+ *
+ * The default value
+ */
+static VALUE default_value(VALUE self)
+{
+  xmlAttributePtr node;
+  Data_Get_Struct(self, xmlAttribute, node);
+
+  if(node->defaultValue) return NOKOGIRI_STR_NEW2(node->defaultValue);
+  return Qnil;
+}
+
+/*
+ * call-seq:
+ *  enumeration
+ *
+ * An enumeration of possible values
+ */
+static VALUE enumeration(VALUE self)
+{
+  xmlAttributePtr node;
+  Data_Get_Struct(self, xmlAttribute, node);
+
+  VALUE list = rb_ary_new();
+  xmlEnumerationPtr enm = node->tree;
+
+  while(enm) {
+    rb_ary_push(list, NOKOGIRI_STR_NEW2(enm->name));
+    enm = enm->next;
+  }
+
+  return list;
+}
+
+VALUE cNokogiriXmlAttributeDecl;
+
+void init_xml_attribute_decl()
+{
+  VALUE nokogiri = rb_define_module("Nokogiri");
+  VALUE xml = rb_define_module_under(nokogiri, "XML");
+  VALUE node = rb_define_class_under(xml, "Node", rb_cObject);
+  VALUE klass = rb_define_class_under(xml, "AttributeDecl", node);
+
+  cNokogiriXmlAttributeDecl = klass;
+
+  rb_define_method(klass, "attribute_type", attribute_type, 0);
+  rb_define_method(klass, "default", default_value, 0);
+  rb_define_method(klass, "enumeration", enumeration, 0);
+}

data/ext/nokogiri/xml_attribute_decl.h

@@ -0,0 +1,9 @@
+#ifndef NOKOGIRI_XML_ATTRIBUTE_DECL
+#define NOKOGIRI_XML_ATTRIBUTE_DECL
+
+#include <nokogiri.h>
+
+void init_xml_attribute_decl();
+
+extern VALUE cNokogiriXmlAttributeDecl;
+#endif

data/ext/nokogiri/xml_cdata.c

@@ -0,0 +1,54 @@
+#include <xml_cdata.h>
+
+/*
+ * call-seq:
+ *  new(document, content)
+ *
+ * Create a new CDATA element on the +document+ with +content+
+ */
+static VALUE new(int argc, VALUE *argv, VALUE klass)
+{
+  xmlDocPtr xml_doc;
+  VALUE doc;
+  VALUE content;
+  VALUE rest;
+
+  rb_scan_args(argc, argv, "2*", &doc, &content, &rest);
+
+  Data_Get_Struct(doc, xmlDoc, xml_doc);
+
+  xmlNodePtr node = xmlNewCDataBlock(
+      xml_doc->doc,
+      NIL_P(content) ? NULL : (const xmlChar *)StringValuePtr(content),
+      NIL_P(content) ? 0 : (int)RSTRING_LEN(content)
+  );
+
+  NOKOGIRI_ROOT_NODE(node);
+
+  VALUE rb_node = Nokogiri_wrap_xml_node(klass, node);
+  rb_obj_call_init(rb_node, argc, argv);
+
+  if(rb_block_given_p()) rb_yield(rb_node);
+
+  return rb_node;
+}
+
+VALUE cNokogiriXmlCData;
+void init_xml_cdata()
+{
+  VALUE nokogiri = rb_define_module("Nokogiri");
+  VALUE xml = rb_define_module_under(nokogiri, "XML");
+  VALUE node = rb_define_class_under(xml, "Node", rb_cObject);
+  VALUE char_data = rb_define_class_under(xml, "CharacterData", node);
+  VALUE text = rb_define_class_under(xml, "Text", char_data);
+
+  /*
+   * CData represents a CData node in an xml document.
+   */
+  VALUE klass = rb_define_class_under(xml, "CDATA", text);
+
+
+  cNokogiriXmlCData = klass;
+
+  rb_define_singleton_method(klass, "new", new, -1);
+}

data/ext/nokogiri/xml_cdata.h

@@ -0,0 +1,9 @@
+#ifndef NOKOGIRI_XML_CDATA
+#define NOKOGIRI_XML_CDATA
+
+#include <nokogiri.h>
+
+void init_xml_cdata();
+
+extern VALUE cNokogiriXmlCData;
+#endif

data/ext/nokogiri/xml_comment.c

@@ -0,0 +1,52 @@
+#include <xml_comment.h>
+
+/*
+ * call-seq:
+ *  new(document, content)
+ *
+ * Create a new Comment element on the +document+ with +content+
+ */
+static VALUE new(int argc, VALUE *argv, VALUE klass)
+{
+  xmlDocPtr xml_doc;
+  VALUE document;
+  VALUE content;
+  VALUE rest;
+
+  rb_scan_args(argc, argv, "2*", &document, &content, &rest);
+
+  Data_Get_Struct(document, xmlDoc, xml_doc);
+
+  xmlNodePtr node = xmlNewDocComment(
+      xml_doc,
+      (const xmlChar *)StringValuePtr(content)
+  );
+
+  VALUE rb_node = Nokogiri_wrap_xml_node(klass, node);
+  rb_obj_call_init(rb_node, argc, argv);
+
+  NOKOGIRI_ROOT_NODE(node);
+
+  if(rb_block_given_p()) rb_yield(rb_node);
+
+  return rb_node;
+}
+
+VALUE cNokogiriXmlComment;
+void init_xml_comment()
+{
+  VALUE nokogiri = rb_define_module("Nokogiri");
+  VALUE xml = rb_define_module_under(nokogiri, "XML");
+  VALUE node = rb_define_class_under(xml, "Node", rb_cObject);
+  VALUE char_data = rb_define_class_under(xml, "CharacterData", node);
+
+  /*
+   * Comment represents a comment node in an xml document.
+   */
+  VALUE klass = rb_define_class_under(xml, "Comment", char_data);
+
+
+  cNokogiriXmlComment = klass;
+
+  rb_define_singleton_method(klass, "new", new, -1);
+}

data/ext/nokogiri/xml_comment.h

@@ -0,0 +1,9 @@
+#ifndef NOKOGIRI_XML_COMMENT
+#define NOKOGIRI_XML_COMMENT
+
+#include <nokogiri.h>
+
+void init_xml_comment();
+
+extern VALUE cNokogiriXmlComment;
+#endif

data/ext/nokogiri/xml_document.c

@@ -0,0 +1,386 @@
+#include <xml_document.h>
+
+static int dealloc_node_i(xmlNodePtr key, xmlNodePtr node, xmlDocPtr doc)
+{
+  switch(node->type) {
+  case XML_ATTRIBUTE_NODE:
+    xmlFreePropList((xmlAttrPtr)node);
+    break;
+  default:
+    if(node->parent == NULL) {
+      xmlAddChild((xmlNodePtr)doc, node);
+    }
+  }
+  return ST_CONTINUE;
+}
+
+static void dealloc(xmlDocPtr doc)
+{
+  NOKOGIRI_DEBUG_START(doc);
+
+  st_table *node_hash = DOC_UNLINKED_NODE_HASH(doc);
+
+  xmlDeregisterNodeFunc func = xmlDeregisterNodeDefault(NULL);
+
+  st_foreach(node_hash, dealloc_node_i, (st_data_t)doc);
+  st_free_table(node_hash);
+
+  free(doc->_private);
+  doc->_private = NULL;
+  xmlFreeDoc(doc);
+
+  xmlDeregisterNodeDefault(func);
+
+  NOKOGIRI_DEBUG_END(doc);
+}
+
+static void recursively_remove_namespaces_from_node(xmlNodePtr node)
+{
+  xmlNodePtr child ;
+
+  xmlSetNs(node, NULL);
+
+  for (child = node->children ; child ; child = child->next)
+    recursively_remove_namespaces_from_node(child);
+}
+
+/*
+ * call-seq:
+ *  url
+ *
+ * Get the url name for this document.
+ */
+static VALUE url(VALUE self)
+{
+  xmlDocPtr doc;
+  Data_Get_Struct(self, xmlDoc, doc);
+
+  if(doc->URL) return NOKOGIRI_STR_NEW2(doc->URL);
+
+  return Qnil;
+}
+
+/*
+ * call-seq:
+ *  root=
+ *
+ * Set the root element on this document
+ */
+static VALUE set_root(VALUE self, VALUE root)
+{
+  xmlDocPtr doc;
+  xmlNodePtr new_root;
+
+  Data_Get_Struct(self, xmlDoc, doc);
+  Data_Get_Struct(root, xmlNode, new_root);
+
+  xmlNodePtr old_root = NULL;
+
+  /* If the new root's document is not the same as the current document,
+   * then we need to dup the node in to this document. */
+  if(new_root->doc != doc) {
+    old_root = xmlDocGetRootElement(doc);
+    if (!(new_root = xmlDocCopyNode(new_root, doc, 1))) {
+      rb_raise(rb_eRuntimeError, "Could not reparent node (xmlDocCopyNode)");
+    }
+  }
+
+  xmlDocSetRootElement(doc, new_root);
+  if(old_root) NOKOGIRI_ROOT_NODE(old_root);
+  return root;
+}
+
+/*
+ * call-seq:
+ *  root
+ *
+ * Get the root node for this document.
+ */
+static VALUE root(VALUE self)
+{
+  xmlDocPtr doc;
+  Data_Get_Struct(self, xmlDoc, doc);
+
+  xmlNodePtr root = xmlDocGetRootElement(doc);
+
+  if(!root) return Qnil;
+  return Nokogiri_wrap_xml_node(Qnil, root) ;
+}
+
+/*
+ * call-seq:
+ *  encoding= encoding
+ *
+ * Set the encoding string for this Document
+ */
+static VALUE set_encoding(VALUE self, VALUE encoding)
+{
+  xmlDocPtr doc;
+  Data_Get_Struct(self, xmlDoc, doc);
+
+  doc->encoding = xmlStrdup((xmlChar *)StringValuePtr(encoding));
+
+  return encoding;
+}
+
+/*
+ * call-seq:
+ *  encoding
+ *
+ * Get the encoding for this Document
+ */
+static VALUE encoding(VALUE self)
+{
+  xmlDocPtr doc;
+  Data_Get_Struct(self, xmlDoc, doc);
+
+  if(!doc->encoding) return Qnil;
+  return NOKOGIRI_STR_NEW2(doc->encoding);
+}
+
+/*
+ * call-seq:
+ *  version
+ *
+ * Get the XML version for this Document
+ */
+static VALUE version(VALUE self)
+{
+  xmlDocPtr doc;
+  Data_Get_Struct(self, xmlDoc, doc);
+
+  if(!doc->version) return Qnil;
+  return NOKOGIRI_STR_NEW2(doc->version);
+}
+
+/*
+ * call-seq:
+ *  read_io(io, url, encoding, options)
+ *
+ * Create a new document from an IO object
+ */
+static VALUE read_io( VALUE klass,
+                      VALUE io,
+                      VALUE url,
+                      VALUE encoding,
+                      VALUE options )
+{
+  const char * c_url    = NIL_P(url)      ? NULL : StringValuePtr(url);
+  const char * c_enc    = NIL_P(encoding) ? NULL : StringValuePtr(encoding);
+  VALUE error_list      = rb_ary_new();
+
+  xmlResetLastError();
+  xmlSetStructuredErrorFunc((void *)error_list, Nokogiri_error_array_pusher);
+
+  xmlDocPtr doc = xmlReadIO(
+      (xmlInputReadCallback)io_read_callback,
+      (xmlInputCloseCallback)io_close_callback,
+      (void *)io,
+      c_url,
+      c_enc,
+      (int)NUM2INT(options)
+  );
+  xmlSetStructuredErrorFunc(NULL, NULL);
+
+  if(doc == NULL) {
+    xmlFreeDoc(doc);
+
+    xmlErrorPtr error = xmlGetLastError();
+    if(error)
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+    else
+      rb_raise(rb_eRuntimeError, "Could not parse document");
+
+    return Qnil;
+  }
+
+  VALUE document = Nokogiri_wrap_xml_document(klass, doc);
+  rb_iv_set(document, "@errors", error_list);
+  return document;
+}
+
+/*
+ * call-seq:
+ *  read_memory(string, url, encoding, options)
+ *
+ * Create a new document from a String
+ */
+static VALUE read_memory( VALUE klass,
+                          VALUE string,
+                          VALUE url,
+                          VALUE encoding,
+                          VALUE options )
+{
+  const char * c_buffer = StringValuePtr(string);
+  const char * c_url    = NIL_P(url)      ? NULL : StringValuePtr(url);
+  const char * c_enc    = NIL_P(encoding) ? NULL : StringValuePtr(encoding);
+  int len               = RSTRING_LEN(string);
+  VALUE error_list      = rb_ary_new();
+
+  xmlResetLastError();
+  xmlSetStructuredErrorFunc((void *)error_list, Nokogiri_error_array_pusher);
+  xmlDocPtr doc = xmlReadMemory(c_buffer, len, c_url, c_enc, (int)NUM2INT(options));
+  xmlSetStructuredErrorFunc(NULL, NULL);
+
+  if(doc == NULL) {
+    xmlFreeDoc(doc);
+
+    xmlErrorPtr error = xmlGetLastError();
+    if(error)
+      rb_exc_raise(Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error));
+    else
+      rb_raise(rb_eRuntimeError, "Could not parse document");
+
+    return Qnil;
+  }
+
+  VALUE document = Nokogiri_wrap_xml_document(klass, doc);
+  rb_iv_set(document, "@errors", error_list);
+  return document;
+}
+
+/*
+ * call-seq:
+ *  dup
+ *
+ * Copy this Document.  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(int argc, VALUE *argv, VALUE self)
+{
+  VALUE level;
+
+  if(rb_scan_args(argc, argv, "01", &level) == 0)
+    level = INT2NUM((long)1);
+
+  xmlDocPtr doc, dup;
+  Data_Get_Struct(self, xmlDoc, doc);
+
+  dup = xmlCopyDoc(doc, (int)NUM2INT(level));
+  if(dup == NULL) return Qnil;
+
+  dup->type = doc->type;
+  return Nokogiri_wrap_xml_document(rb_obj_class(self), dup);
+}
+
+/*
+ * call-seq:
+ *  new(version = default)
+ *
+ * Create a new document with +version+ (defaults to "1.0")
+ */
+static VALUE new(int argc, VALUE *argv, VALUE klass)
+{
+  VALUE version, rest, rb_doc ;
+
+  rb_scan_args(argc, argv, "0*", &rest);
+  version = rb_ary_entry(rest, (long)0);
+  if (NIL_P(version)) version = rb_str_new2("1.0");
+
+  xmlDocPtr doc = xmlNewDoc((xmlChar *)StringValuePtr(version));
+  rb_doc = Nokogiri_wrap_xml_document(klass, doc);
+  rb_obj_call_init(rb_doc, argc, argv);
+  return rb_doc ;
+}
+
+/*
+ *  call-seq:
+ *    remove_namespaces!
+ *
+ *  Remove all namespaces from all nodes in the document.
+ *
+ *  This could be useful for developers who either don't understand namespaces
+ *  or don't care about them.
+ *
+ *  The following example shows a use case, and you can decide for yourself
+ *  whether this is a good thing or not:
+ *
+ *    doc = Nokogiri::XML <<-EOXML
+ *       <root>
+ *         <car xmlns:part="http://general-motors.com/">
+ *           <part:tire>Michelin Model XGV</part:tire>
+ *         </car>
+ *         <bicycle xmlns:part="http://schwinn.com/">
+ *           <part:tire>I'm a bicycle tire!</part:tire>
+ *         </bicycle>
+ *       </root>
+ *       EOXML
+ *    
+ *    doc.xpath("//tire").to_s # => ""
+ *    doc.xpath("//part:tire", "part" => "http://general-motors.com/").to_s # => "<part:tire>Michelin Model XGV</part:tire>"
+ *    doc.xpath("//part:tire", "part" => "http://schwinn.com/").to_s # => "<part:tire>I'm a bicycle tire!</part:tire>"
+ *    
+ *    doc.remove_namespaces!
+ *    
+ *    doc.xpath("//tire").to_s # => "<tire>Michelin Model XGV</tire><tire>I'm a bicycle tire!</tire>"
+ *    doc.xpath("//part:tire", "part" => "http://general-motors.com/").to_s # => ""
+ *    doc.xpath("//part:tire", "part" => "http://schwinn.com/").to_s # => ""
+ *
+ *  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/
+ */
+VALUE remove_namespaces_bang(VALUE self)
+{
+  xmlDocPtr doc ;
+  Data_Get_Struct(self, xmlDoc, doc);
+
+  recursively_remove_namespaces_from_node((xmlNodePtr)doc);
+  return self;
+}
+
+
+VALUE cNokogiriXmlDocument ;
+void init_xml_document()
+{
+  VALUE nokogiri  = rb_define_module("Nokogiri");
+  VALUE xml       = rb_define_module_under(nokogiri, "XML");
+  VALUE node      = rb_define_class_under(xml, "Node", rb_cObject);
+
+  /*
+   * Nokogiri::XML::Document wraps an xml document.
+   */
+  VALUE klass = rb_define_class_under(xml, "Document", node);
+
+  cNokogiriXmlDocument = klass;
+
+  rb_define_singleton_method(klass, "read_memory", read_memory, 4);
+  rb_define_singleton_method(klass, "read_io", read_io, 4);
+  rb_define_singleton_method(klass, "new", new, -1);
+
+  rb_define_method(klass, "root", root, 0);
+  rb_define_method(klass, "root=", set_root, 1);
+  rb_define_method(klass, "encoding", encoding, 0);
+  rb_define_method(klass, "encoding=", set_encoding, 1);
+  rb_define_method(klass, "version", version, 0);
+  rb_define_method(klass, "dup", duplicate_node, -1);
+  rb_define_method(klass, "url", url, 0);
+  rb_define_method(klass, "remove_namespaces!", remove_namespaces_bang, 0);
+}
+
+
+/* this takes klass as a param because it's used for HtmlDocument, too. */
+VALUE Nokogiri_wrap_xml_document(VALUE klass, xmlDocPtr doc)
+{
+  nokogiriTuplePtr tuple = (nokogiriTuplePtr)malloc(sizeof(nokogiriTuple));
+
+  VALUE rb_doc = Data_Wrap_Struct(
+      klass ? klass : cNokogiriXmlDocument,
+      0,
+      dealloc,
+      doc
+  );
+
+  VALUE cache = rb_ary_new();
+  rb_iv_set(rb_doc, "@decorators", Qnil);
+  rb_iv_set(rb_doc, "@node_cache", cache);
+
+  tuple->doc = (void *)rb_doc;
+  tuple->unlinkedNodes = st_init_numtable_with_size(128);
+  tuple->node_cache = cache;
+  doc->_private = tuple ;
+
+  rb_obj_call_init(rb_doc, 0, NULL);
+
+  return rb_doc ;
+}