nokogiri 1.13.6 → 1.14.2

data/ext/nokogiri/xml_node_set.c

@@ -88,13 +88,11 @@ deallocate(xmlNodeSetPtr node_set)
    * For reasons outlined in xml_namespace.c, here we reproduce xmlXPathFreeNodeSet() except for the
    * offending call to xmlXPathNodeSetFreeNs().
    */
-  NOKOGIRI_DEBUG_START(node_set) ;
   if (node_set->nodeTab != NULL) {
     xmlFree(node_set->nodeTab);
   }
 
   xmlFree(node_set);
-  NOKOGIRI_DEBUG_END(node_set) ;
 }
 
 
@@ -156,7 +154,7 @@ push(VALUE self, VALUE rb_node)
   Check_Node_Set_Node_Type(rb_node);
 
   Data_Get_Struct(self, xmlNodeSet, node_set);
-  Data_Get_Struct(rb_node, xmlNode, node);
+  Noko_Node_Get_Struct(rb_node, xmlNode, node);
 
   xmlXPathNodeSetAdd(node_set, node);
 
@@ -179,7 +177,7 @@ delete (VALUE self, VALUE rb_node)
   Check_Node_Set_Node_Type(rb_node);
 
   Data_Get_Struct(self, xmlNodeSet, node_set);
-  Data_Get_Struct(rb_node, xmlNode, node);
+  Noko_Node_Get_Struct(rb_node, xmlNode, node);
 
   if (xmlXPathNodeSetContains(node_set, node)) {
     xpath_node_set_del(node_set, node);
@@ -228,7 +226,7 @@ include_eh(VALUE self, VALUE rb_node)
   Check_Node_Set_Node_Type(rb_node);
 
   Data_Get_Struct(self, xmlNodeSet, node_set);
-  Data_Get_Struct(rb_node, xmlNode, node);
+  Noko_Node_Get_Struct(rb_node, xmlNode, node);
 
   return (xmlXPathNodeSetContains(node_set, node) ? Qtrue : Qfalse);
 }
@@ -430,7 +428,7 @@ unlink_nodeset(VALUE self)
       xmlNodePtr node_ptr;
       node = noko_xml_node_wrap(Qnil, node_set->nodeTab[j]);
       rb_funcall(node, rb_intern("unlink"), 0); /* modifies the C struct out from under the object */
-      Data_Get_Struct(node, xmlNode, node_ptr);
+      Noko_Node_Get_Struct(node, xmlNode, node_ptr);
       node_set->nodeTab[j] = node_ptr ;
     }
   }

data/ext/nokogiri/xml_processing_instruction.c

@@ -41,7 +41,7 @@ new (int argc, VALUE *argv, VALUE klass)
 }
 
 void
-noko_init_xml_processing_instruction()
+noko_init_xml_processing_instruction(void)
 {
   assert(cNokogiriXmlNode);
   /*

data/ext/nokogiri/xml_reader.c

@@ -5,9 +5,7 @@ VALUE cNokogiriXmlReader;
 static void
 dealloc(xmlTextReaderPtr reader)
 {
-  NOKOGIRI_DEBUG_START(reader);
   xmlFreeTextReader(reader);
-  NOKOGIRI_DEBUG_END(reader);
 }
 
 static int
@@ -31,6 +29,7 @@ has_attributes(xmlTextReaderPtr reader)
   return (0);
 }
 
+// TODO: merge this function into the `namespaces` method implementation
 static void
 Nokogiri_xml_node_namespaces(xmlNodePtr node, VALUE attr_hash)
 {
@@ -125,32 +124,47 @@ attributes_eh(VALUE self)
  * Get a hash of namespaces for this Node
  */
 static VALUE
-namespaces(VALUE self)
+rb_xml_reader_namespaces(VALUE rb_reader)
 {
-  xmlTextReaderPtr reader;
-  xmlNodePtr ptr;
-  VALUE attr ;
-
-  Data_Get_Struct(self, xmlTextReader, reader);
+  VALUE rb_namespaces = rb_hash_new() ;
+  xmlTextReaderPtr c_reader;
+  xmlNodePtr c_node;
+  VALUE rb_errors;
 
-  attr = rb_hash_new() ;
+  Data_Get_Struct(rb_reader, xmlTextReader, c_reader);
 
-  if (! has_attributes(reader)) {
-    return attr ;
+  if (! has_attributes(c_reader)) {
+    return rb_namespaces ;
   }
 
-  ptr = xmlTextReaderExpand(reader);
-  if (ptr == NULL) { return Qnil; }
+  rb_errors = rb_funcall(rb_reader, rb_intern("errors"), 0);
+
+  xmlSetStructuredErrorFunc((void *)rb_errors, Nokogiri_error_array_pusher);
+  c_node = xmlTextReaderExpand(c_reader);
+  xmlSetStructuredErrorFunc(NULL, NULL);
+
+  if (c_node == NULL) {
+    if (RARRAY_LEN(rb_errors) > 0) {
+      VALUE rb_error = rb_ary_entry(rb_errors, 0);
+      VALUE exception_message = rb_funcall(rb_error, rb_intern("to_s"), 0);
+      rb_exc_raise(rb_class_new_instance(1, &exception_message, cNokogiriXmlSyntaxError));
+    }
+    return Qnil;
+  }
 
-  Nokogiri_xml_node_namespaces(ptr, attr);
+  Nokogiri_xml_node_namespaces(c_node, rb_namespaces);
 
-  return attr ;
+  return rb_namespaces ;
 }
 
 /*
   :call-seq: attribute_nodes() → Array<Nokogiri::XML::Attr>
 
-  Get the attributes of the current node as an Array of Attr
+  Get the attributes of the current node as an Array of XML:Attr
+
+  ⚠ This method is deprecated and unsafe to use. It will be removed in a future version of Nokogiri.
+
+  See related: #attribute_hash, #attributes
  */
 static VALUE
 rb_xml_reader_attribute_nodes(VALUE rb_reader)
@@ -160,6 +174,10 @@ rb_xml_reader_attribute_nodes(VALUE rb_reader)
   VALUE attr_nodes;
   int j;
 
+  // TODO: deprecated, remove in Nokogiri v1.15, see https://github.com/sparklemotion/nokogiri/issues/2598
+  // After removal, we can also remove all the "node_has_a_document" special handling from xml_node.c
+  NOKO_WARN_DEPRECATION("Reader#attribute_nodes is deprecated and will be removed in a future version of Nokogiri. Please use Reader#attribute_hash instead.");
+
   Data_Get_Struct(rb_reader, xmlTextReader, c_reader);
 
   if (! has_attributes(c_reader)) {
@@ -181,6 +199,62 @@ rb_xml_reader_attribute_nodes(VALUE rb_reader)
   return attr_nodes;
 }
 
+/*
+  :call-seq: attribute_hash() → Hash<String ⇒ String>
+
+  Get the attributes of the current node as a Hash of names and values.
+
+  See related: #attributes and #namespaces
+ */
+static VALUE
+rb_xml_reader_attribute_hash(VALUE rb_reader)
+{
+  VALUE rb_attributes = rb_hash_new();
+  xmlTextReaderPtr c_reader;
+  xmlNodePtr c_node;
+  xmlAttrPtr c_property;
+  VALUE rb_errors;
+
+  Data_Get_Struct(rb_reader, xmlTextReader, c_reader);
+
+  if (!has_attributes(c_reader)) {
+    return rb_attributes;
+  }
+
+  rb_errors = rb_funcall(rb_reader, rb_intern("errors"), 0);
+
+  xmlSetStructuredErrorFunc((void *)rb_errors, Nokogiri_error_array_pusher);
+  c_node = xmlTextReaderExpand(c_reader);
+  xmlSetStructuredErrorFunc(NULL, NULL);
+
+  if (c_node == NULL) {
+    if (RARRAY_LEN(rb_errors) > 0) {
+      VALUE rb_error = rb_ary_entry(rb_errors, 0);
+      VALUE exception_message = rb_funcall(rb_error, rb_intern("to_s"), 0);
+      rb_exc_raise(rb_class_new_instance(1, &exception_message, cNokogiriXmlSyntaxError));
+    }
+    return Qnil;
+  }
+
+  c_property = c_node->properties;
+  while (c_property != NULL) {
+    VALUE rb_name = NOKOGIRI_STR_NEW2(c_property->name);
+    VALUE rb_value = Qnil;
+    xmlChar *c_value = xmlNodeGetContent((xmlNode *)c_property);
+
+    if (c_value) {
+      rb_value = NOKOGIRI_STR_NEW2(c_value);
+      xmlFree(c_value);
+    }
+
+    rb_hash_aset(rb_attributes, rb_name, rb_value);
+
+    c_property = c_property->next;
+  }
+
+  return rb_attributes;
+}
+
 /*
  * call-seq:
  *   attribute_at(index)
@@ -252,7 +326,7 @@ attribute_count(VALUE self)
   count = xmlTextReaderAttributeCount(reader);
   if (count == -1) { return Qnil; }
 
-  return INT2NUM((long)count);
+  return INT2NUM(count);
 }
 
 /*
@@ -271,7 +345,7 @@ depth(VALUE self)
   depth = xmlTextReaderDepth(reader);
   if (depth == -1) { return Qnil; }
 
-  return INT2NUM((long)depth);
+  return INT2NUM(depth);
 }
 
 /*
@@ -444,7 +518,7 @@ state(VALUE self)
 {
   xmlTextReaderPtr reader;
   Data_Get_Struct(self, xmlTextReader, reader);
-  return INT2NUM((long)xmlTextReaderReadState(reader));
+  return INT2NUM(xmlTextReaderReadState(reader));
 }
 
 /*
@@ -458,7 +532,7 @@ node_type(VALUE self)
 {
   xmlTextReaderPtr reader;
   Data_Get_Struct(self, xmlTextReader, reader);
-  return INT2NUM((long)xmlTextReaderNodeType(reader));
+  return INT2NUM(xmlTextReaderNodeType(reader));
 }
 
 /*
@@ -678,7 +752,7 @@ rb_xml_reader_encoding(VALUE rb_reader)
 }
 
 void
-noko_init_xml_reader()
+noko_init_xml_reader(void)
 {
   /*
    * The Reader parser allows you to effectively pull parse an XML document.
@@ -696,6 +770,7 @@ noko_init_xml_reader()
   rb_define_method(cNokogiriXmlReader, "attribute_at", attribute_at, 1);
   rb_define_method(cNokogiriXmlReader, "attribute_count", attribute_count, 0);
   rb_define_method(cNokogiriXmlReader, "attribute_nodes", rb_xml_reader_attribute_nodes, 0);
+  rb_define_method(cNokogiriXmlReader, "attribute_hash", rb_xml_reader_attribute_hash, 0);
   rb_define_method(cNokogiriXmlReader, "attributes?", attributes_eh, 0);
   rb_define_method(cNokogiriXmlReader, "base_uri", rb_xml_reader_base_uri, 0);
   rb_define_method(cNokogiriXmlReader, "default?", default_eh, 0);
@@ -707,7 +782,7 @@ noko_init_xml_reader()
   rb_define_method(cNokogiriXmlReader, "local_name", local_name, 0);
   rb_define_method(cNokogiriXmlReader, "name", name, 0);
   rb_define_method(cNokogiriXmlReader, "namespace_uri", namespace_uri, 0);
-  rb_define_method(cNokogiriXmlReader, "namespaces", namespaces, 0);
+  rb_define_method(cNokogiriXmlReader, "namespaces", rb_xml_reader_namespaces, 0);
   rb_define_method(cNokogiriXmlReader, "node_type", node_type, 0);
   rb_define_method(cNokogiriXmlReader, "outer_xml", outer_xml, 0);
   rb_define_method(cNokogiriXmlReader, "prefix", prefix, 0);

data/ext/nokogiri/xml_relax_ng.c

@@ -5,9 +5,7 @@ VALUE cNokogiriXmlRelaxNG;
 static void
 dealloc(xmlRelaxNGPtr schema)
 {
-  NOKOGIRI_DEBUG_START(schema);
   xmlRelaxNGFree(schema);
-  NOKOGIRI_DEBUG_END(schema);
 }
 
 /*
@@ -173,7 +171,7 @@ from_document(int argc, VALUE *argv, VALUE klass)
 }
 
 void
-noko_init_xml_relax_ng()
+noko_init_xml_relax_ng(void)
 {
   assert(cNokogiriXmlSchema);
   cNokogiriXmlRelaxNG = rb_define_class_under(mNokogiriXml, "RelaxNG", cNokogiriXmlSchema);

data/ext/nokogiri/xml_sax_parser.c

@@ -195,40 +195,48 @@ comment_func(void *ctx, const xmlChar *value)
   rb_funcall(doc, id_comment, 1, str);
 }
 
+PRINTFLIKE_DECL(2, 3)
 static void
 warning_func(void *ctx, const char *msg, ...)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
-  char *message;
-  VALUE ruby_message;
+  VALUE rb_message;
 
+#ifdef TRUFFLERUBY_NOKOGIRI_SYSTEM_LIBRARIES
+  /* It is not currently possible to pass var args from native
+     functions to sulong, so we work around the issue here. */
+  rb_message = rb_sprintf("warning_func: %s", msg);
+#else
   va_list args;
   va_start(args, msg);
-  vasprintf(&message, msg, args);
+  rb_message = rb_vsprintf(msg, args);
   va_end(args);
+#endif
 
-  ruby_message = NOKOGIRI_STR_NEW2(message);
-  free(message);
-  rb_funcall(doc, id_warning, 1, ruby_message);
+  rb_funcall(doc, id_warning, 1, rb_message);
 }
 
+PRINTFLIKE_DECL(2, 3)
 static void
 error_func(void *ctx, const char *msg, ...)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
-  char *message;
-  VALUE ruby_message;
+  VALUE rb_message;
 
+#ifdef TRUFFLERUBY_NOKOGIRI_SYSTEM_LIBRARIES
+  /* It is not currently possible to pass var args from native
+     functions to sulong, so we work around the issue here. */
+  rb_message = rb_sprintf("error_func: %s", msg);
+#else
   va_list args;
   va_start(args, msg);
-  vasprintf(&message, msg, args);
+  rb_message = rb_vsprintf(msg, args);
   va_end(args);
+#endif
 
-  ruby_message = NOKOGIRI_STR_NEW2(message);
-  free(message);
-  rb_funcall(doc, id_error, 1, ruby_message);
+  rb_funcall(doc, id_error, 1, rb_message);
 }
 
 static void
@@ -260,15 +268,13 @@ processing_instruction(void *ctx, const xmlChar *name, const xmlChar *content)
 static void
 deallocate(xmlSAXHandlerPtr handler)
 {
-  NOKOGIRI_DEBUG_START(handler);
-  free(handler);
-  NOKOGIRI_DEBUG_END(handler);
+  ruby_xfree(handler);
 }
 
 static VALUE
 allocate(VALUE klass)
 {
-  xmlSAXHandlerPtr handler = calloc((size_t)1, sizeof(xmlSAXHandler));
+  xmlSAXHandlerPtr handler = ruby_xcalloc((size_t)1, sizeof(xmlSAXHandler));
 
   handler->startDocument = start_document;
   handler->endDocument = end_document;
@@ -288,7 +294,7 @@ allocate(VALUE klass)
 }
 
 void
-noko_init_xml_sax_parser()
+noko_init_xml_sax_parser(void)
 {
   cNokogiriXmlSaxParser = rb_define_class_under(mNokogiriXmlSax, "Parser", rb_cObject);
 

data/ext/nokogiri/xml_sax_parser_context.c

@@ -7,13 +7,8 @@ static ID id_read;
 static void
 deallocate(xmlParserCtxtPtr ctxt)
 {
-  NOKOGIRI_DEBUG_START(ctxt);
-
   ctxt->sax = NULL;
-
   xmlFreeParserCtxt(ctxt);
-
-  NOKOGIRI_DEBUG_END(ctxt);
 }
 
 /*
@@ -266,7 +261,7 @@ get_recovery(VALUE self)
 }
 
 void
-noko_init_xml_sax_parser_context()
+noko_init_xml_sax_parser_context(void)
 {
   cNokogiriXmlSaxParserContext = rb_define_class_under(mNokogiriXmlSax, "ParserContext", rb_cObject);
 

data/ext/nokogiri/xml_sax_push_parser.c

@@ -5,12 +5,10 @@ VALUE cNokogiriXmlSaxPushParser ;
 static void
 deallocate(xmlParserCtxtPtr ctx)
 {
-  NOKOGIRI_DEBUG_START(ctx);
   if (ctx != NULL) {
     NOKOGIRI_SAX_TUPLE_DESTROY(ctx->userData);
     xmlFreeParserCtxt(ctx);
   }
-  NOKOGIRI_DEBUG_END(ctx);
 }
 
 static VALUE
@@ -152,7 +150,7 @@ set_replace_entities(VALUE self, VALUE value)
 }
 
 void
-noko_init_xml_sax_push_parser()
+noko_init_xml_sax_push_parser(void)
 {
   cNokogiriXmlSaxPushParser = rb_define_class_under(mNokogiriXmlSax, "PushParser", rb_cObject);
 

data/ext/nokogiri/xml_schema.c

@@ -5,9 +5,7 @@ VALUE cNokogiriXmlSchema;
 static void
 dealloc(xmlSchemaPtr schema)
 {
-  NOKOGIRI_DEBUG_START(schema);
   xmlSchemaFree(schema);
-  NOKOGIRI_DEBUG_END(schema);
 }
 
 /*
@@ -25,7 +23,7 @@ validate_document(VALUE self, VALUE document)
   VALUE errors;
 
   Data_Get_Struct(self, xmlSchema, schema);
-  Data_Get_Struct(document, xmlDoc, doc);
+  Noko_Node_Get_Struct(document, xmlDoc, doc);
 
   errors = rb_ary_new();
 
@@ -179,7 +177,7 @@ has_blank_nodes_p(VALUE cache)
   for (i = 0; i < RARRAY_LEN(cache); i++) {
     xmlNodePtr node;
     VALUE element = rb_ary_entry(cache, i);
-    Data_Get_Struct(element, xmlNode, node);
+    Noko_Node_Get_Struct(element, xmlNode, node);
     if (xmlIsBlankNode(node)) {
       return 1;
     }
@@ -210,7 +208,7 @@ from_document(int argc, VALUE *argv, VALUE klass)
 
   scanned_args = rb_scan_args(argc, argv, "11", &document, &parse_options);
 
-  Data_Get_Struct(document, xmlDoc, doc);
+  Noko_Node_Get_Struct(document, xmlDoc, doc);
   doc = doc->doc; /* In case someone passes us a node. ugh. */
 
   if (scanned_args == 1) {
@@ -270,7 +268,7 @@ from_document(int argc, VALUE *argv, VALUE klass)
 }
 
 void
-noko_init_xml_schema()
+noko_init_xml_schema(void)
 {
   cNokogiriXmlSchema = rb_define_class_under(mNokogiriXml, "Schema", rb_cObject);
 

data/ext/nokogiri/xml_syntax_error.c

@@ -75,7 +75,7 @@ Nokogiri_wrap_xml_syntax_error(xmlErrorPtr error)
 }
 
 void
-noko_init_xml_syntax_error()
+noko_init_xml_syntax_error(void)
 {
   assert(cNokogiriSyntaxError);
   /*

data/ext/nokogiri/xml_text.c

@@ -20,7 +20,7 @@ new (int argc, VALUE *argv, VALUE klass)
 
   rb_scan_args(argc, argv, "2*", &string, &document, &rest);
 
-  Data_Get_Struct(document, xmlDoc, doc);
+  Noko_Node_Get_Struct(document, xmlDoc, doc);
 
   node = xmlNewText((xmlChar *)StringValueCStr(string));
   node->doc = doc->doc;
@@ -36,7 +36,7 @@ new (int argc, VALUE *argv, VALUE klass)
 }
 
 void
-noko_init_xml_text()
+noko_init_xml_text(void)
 {
   assert(cNokogiriXmlCharacterData);
   /*