nokogiri 1.13.8 → 1.15.7

data/ext/nokogiri/xml_sax_parser_context.c

@@ -5,17 +5,40 @@ VALUE cNokogiriXmlSaxParserContext ;
 static ID id_read;
 
 static void
-deallocate(xmlParserCtxtPtr ctxt)
+xml_sax_parser_context_free(void *data)
 {
-  NOKOGIRI_DEBUG_START(ctxt);
-
+  xmlParserCtxtPtr ctxt = data;
   ctxt->sax = NULL;
-
   xmlFreeParserCtxt(ctxt);
+}
+
+/*
+ *  note that htmlParserCtxtPtr == xmlParserCtxtPtr and xmlFreeParserCtxt() == htmlFreeParserCtxt()
+ *  so we use this type for both XML::SAX::ParserContext and HTML::SAX::ParserContext
+ */
+static const rb_data_type_t xml_sax_parser_context_type = {
+  .wrap_struct_name = "Nokogiri::XML::SAX::ParserContext",
+  .function = {
+    .dfree = xml_sax_parser_context_free,
+  },
+  .flags = RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_WB_PROTECTED,
+};
+
+xmlParserCtxtPtr
+noko_xml_sax_parser_context_unwrap(VALUE rb_context)
+{
+  xmlParserCtxtPtr c_context;
+  TypedData_Get_Struct(rb_context, xmlParserCtxt, &xml_sax_parser_context_type, c_context);
+  return c_context;
+}
 
-  NOKOGIRI_DEBUG_END(ctxt);
+VALUE
+noko_xml_sax_parser_context_wrap(VALUE klass, xmlParserCtxtPtr c_context)
+{
+  return TypedData_Wrap_Struct(klass, &xml_sax_parser_context_type, c_context);
 }
 
+
 /*
  * call-seq:
  *  parse_io(io, encoding)
@@ -41,7 +64,7 @@ parse_io(VALUE klass, VALUE io, VALUE encoding)
     ctxt->sax = NULL;
   }
 
-  return Data_Wrap_Struct(klass, NULL, deallocate, ctxt);
+  return noko_xml_sax_parser_context_wrap(klass, ctxt);
 }
 
 /*
@@ -54,7 +77,13 @@ static VALUE
 parse_file(VALUE klass, VALUE filename)
 {
   xmlParserCtxtPtr ctxt = xmlCreateFileParserCtxt(StringValueCStr(filename));
-  return Data_Wrap_Struct(klass, NULL, deallocate, ctxt);
+
+  if (ctxt->sax) {
+    xmlFree(ctxt->sax);
+    ctxt->sax = NULL;
+  }
+
+  return noko_xml_sax_parser_context_wrap(klass, ctxt);
 }
 
 /*
@@ -81,7 +110,7 @@ parse_memory(VALUE klass, VALUE data)
     ctxt->sax = NULL;
   }
 
-  return Data_Wrap_Struct(klass, NULL, deallocate, ctxt);
+  return noko_xml_sax_parser_context_wrap(klass, ctxt);
 }
 
 static VALUE
@@ -121,13 +150,8 @@ parse_with(VALUE self, VALUE sax_handler)
     rb_raise(rb_eArgError, "argument must be a Nokogiri::XML::SAX::Parser");
   }
 
-  Data_Get_Struct(self, xmlParserCtxt, ctxt);
-  Data_Get_Struct(sax_handler, xmlSAXHandler, sax);
-
-  /* Free the sax handler since we'll assign our own */
-  if (ctxt->sax && ctxt->sax != (xmlSAXHandlerPtr)&xmlDefaultSAXHandler) {
-    xmlFree(ctxt->sax);
-  }
+  ctxt = noko_xml_sax_parser_context_unwrap(self);
+  sax = noko_sax_handler_unwrap(sax_handler);
 
   ctxt->sax = sax;
   ctxt->userData = (void *)NOKOGIRI_SAX_TUPLE_NEW(ctxt, sax_handler);
@@ -149,8 +173,7 @@ parse_with(VALUE self, VALUE sax_handler)
 static VALUE
 set_replace_entities(VALUE self, VALUE value)
 {
-  xmlParserCtxtPtr ctxt;
-  Data_Get_Struct(self, xmlParserCtxt, ctxt);
+  xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
   if (Qfalse == value) {
     ctxt->replaceEntities = 0;
@@ -171,8 +194,7 @@ set_replace_entities(VALUE self, VALUE value)
 static VALUE
 get_replace_entities(VALUE self)
 {
-  xmlParserCtxtPtr ctxt;
-  Data_Get_Struct(self, xmlParserCtxt, ctxt);
+  xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
   if (0 == ctxt->replaceEntities) {
     return Qfalse;
@@ -189,10 +211,8 @@ get_replace_entities(VALUE self)
 static VALUE
 line(VALUE self)
 {
-  xmlParserCtxtPtr ctxt;
   xmlParserInputPtr io;
-
-  Data_Get_Struct(self, xmlParserCtxt, ctxt);
+  xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
   io = ctxt->input;
   if (io) {
@@ -210,11 +230,9 @@ line(VALUE self)
 static VALUE
 column(VALUE self)
 {
-  xmlParserCtxtPtr ctxt;
+  xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
   xmlParserInputPtr io;
 
-  Data_Get_Struct(self, xmlParserCtxt, ctxt);
-
   io = ctxt->input;
   if (io) {
     return INT2NUM(io->col);
@@ -233,8 +251,7 @@ column(VALUE self)
 static VALUE
 set_recovery(VALUE self, VALUE value)
 {
-  xmlParserCtxtPtr ctxt;
-  Data_Get_Struct(self, xmlParserCtxt, ctxt);
+  xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
   if (value == Qfalse) {
     ctxt->recovery = 0;
@@ -255,8 +272,7 @@ set_recovery(VALUE self, VALUE value)
 static VALUE
 get_recovery(VALUE self)
 {
-  xmlParserCtxtPtr ctxt;
-  Data_Get_Struct(self, xmlParserCtxt, ctxt);
+  xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
   if (ctxt->recovery == 0) {
     return Qfalse;
@@ -266,7 +282,7 @@ 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

@@ -3,20 +3,35 @@
 VALUE cNokogiriXmlSaxPushParser ;
 
 static void
-deallocate(xmlParserCtxtPtr ctx)
+xml_sax_push_parser_free(void *data)
 {
-  NOKOGIRI_DEBUG_START(ctx);
+  xmlParserCtxtPtr ctx = data;
   if (ctx != NULL) {
     NOKOGIRI_SAX_TUPLE_DESTROY(ctx->userData);
     xmlFreeParserCtxt(ctx);
   }
-  NOKOGIRI_DEBUG_END(ctx);
 }
 
+static const rb_data_type_t xml_sax_push_parser_type = {
+  .wrap_struct_name = "Nokogiri::XML::SAX::PushParser",
+  .function = {
+    .dfree = xml_sax_push_parser_free,
+  },
+  .flags = RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_WB_PROTECTED,
+};
+
 static VALUE
 allocate(VALUE klass)
 {
-  return Data_Wrap_Struct(klass, NULL, deallocate, NULL);
+  return TypedData_Wrap_Struct(klass, &xml_sax_push_parser_type, NULL);
+}
+
+xmlParserCtxtPtr
+noko_xml_sax_push_parser_unwrap(VALUE rb_parser)
+{
+  xmlParserCtxtPtr c_parser;
+  TypedData_Get_Struct(rb_parser, xmlParserCtxt, &xml_sax_push_parser_type, c_parser);
+  return c_parser;
 }
 
 /*
@@ -33,7 +48,7 @@ native_write(VALUE self, VALUE _chunk, VALUE _last_chunk)
   int size            = 0;
 
 
-  Data_Get_Struct(self, xmlParserCtxt, ctx);
+  ctx = noko_xml_sax_push_parser_unwrap(self);
 
   if (Qnil != _chunk) {
     chunk = StringValuePtr(_chunk);
@@ -65,7 +80,7 @@ initialize_native(VALUE self, VALUE _xml_sax, VALUE _filename)
   const char *filename = NULL;
   xmlParserCtxtPtr ctx;
 
-  Data_Get_Struct(_xml_sax, xmlSAXHandler, sax);
+  sax = noko_sax_handler_unwrap(_xml_sax);
 
   if (_filename != Qnil) { filename = StringValueCStr(_filename); }
 
@@ -91,7 +106,8 @@ static VALUE
 get_options(VALUE self)
 {
   xmlParserCtxtPtr ctx;
-  Data_Get_Struct(self, xmlParserCtxt, ctx);
+
+  ctx = noko_xml_sax_push_parser_unwrap(self);
 
   return INT2NUM(ctx->options);
 }
@@ -100,7 +116,8 @@ static VALUE
 set_options(VALUE self, VALUE options)
 {
   xmlParserCtxtPtr ctx;
-  Data_Get_Struct(self, xmlParserCtxt, ctx);
+
+  ctx = noko_xml_sax_push_parser_unwrap(self);
 
   if (xmlCtxtUseOptions(ctx, (int)NUM2INT(options)) != 0) {
     rb_raise(rb_eRuntimeError, "Cannot set XML parser context options");
@@ -120,7 +137,8 @@ static VALUE
 get_replace_entities(VALUE self)
 {
   xmlParserCtxtPtr ctx;
-  Data_Get_Struct(self, xmlParserCtxt, ctx);
+
+  ctx = noko_xml_sax_push_parser_unwrap(self);
 
   if (0 == ctx->replaceEntities) {
     return Qfalse;
@@ -140,7 +158,8 @@ static VALUE
 set_replace_entities(VALUE self, VALUE value)
 {
   xmlParserCtxtPtr ctx;
-  Data_Get_Struct(self, xmlParserCtxt, ctx);
+
+  ctx = noko_xml_sax_push_parser_unwrap(self);
 
   if (Qfalse == value) {
     ctx->replaceEntities = 0;
@@ -152,7 +171,7 @@ 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

@@ -3,13 +3,20 @@
 VALUE cNokogiriXmlSchema;
 
 static void
-dealloc(xmlSchemaPtr schema)
+xml_schema_deallocate(void *data)
 {
-  NOKOGIRI_DEBUG_START(schema);
+  xmlSchemaPtr schema = data;
   xmlSchemaFree(schema);
-  NOKOGIRI_DEBUG_END(schema);
 }
 
+static const rb_data_type_t xml_schema_type = {
+  .wrap_struct_name = "Nokogiri::XML::Schema",
+  .function = {
+    .dfree = xml_schema_deallocate,
+  },
+  .flags = RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_WB_PROTECTED,
+};
+
 /*
  * call-seq:
  *  validate_document(document)
@@ -24,8 +31,8 @@ validate_document(VALUE self, VALUE document)
   xmlSchemaValidCtxtPtr valid_ctxt;
   VALUE errors;
 
-  Data_Get_Struct(self, xmlSchema, schema);
-  Noko_Node_Get_Struct(document, xmlDoc, doc);
+  TypedData_Get_Struct(self, xmlSchema, &xml_schema_type, schema);
+  doc = noko_xml_document_unwrap(document);
 
   errors = rb_ary_new();
 
@@ -65,7 +72,7 @@ validate_file(VALUE self, VALUE rb_filename)
   const char *filename ;
   VALUE errors;
 
-  Data_Get_Struct(self, xmlSchema, schema);
+  TypedData_Get_Struct(self, xmlSchema, &xml_schema_type, schema);
   filename = (const char *)StringValueCStr(rb_filename) ;
 
   errors = rb_ary_new();
@@ -92,59 +99,53 @@ validate_file(VALUE self, VALUE rb_filename)
   return errors;
 }
 
-/*
- * call-seq:
- *  read_memory(string)
- *
- * Create a new Schema from the contents of +string+
- */
 static VALUE
-read_memory(int argc, VALUE *argv, VALUE klass)
+xml_schema_parse_schema(
+  VALUE klass,
+  xmlSchemaParserCtxtPtr c_parser_context,
+  VALUE rb_parse_options
+)
 {
-  VALUE content;
-  VALUE parse_options;
+  VALUE rb_errors;
   int parse_options_int;
-  xmlSchemaParserCtxtPtr ctx;
-  xmlSchemaPtr schema;
-  VALUE errors;
-  VALUE rb_schema;
-  int scanned_args = 0;
+  xmlSchemaPtr c_schema;
   xmlExternalEntityLoader old_loader = 0;
+  VALUE rb_schema;
 
-  scanned_args = rb_scan_args(argc, argv, "11", &content, &parse_options);
-  if (scanned_args == 1) {
-    parse_options = rb_const_get_at(rb_const_get_at(mNokogiriXml, rb_intern("ParseOptions")), rb_intern("DEFAULT_SCHEMA"));
+  if (NIL_P(rb_parse_options)) {
+    rb_parse_options = rb_const_get_at(
+                         rb_const_get_at(mNokogiriXml, rb_intern("ParseOptions")),
+                         rb_intern("DEFAULT_SCHEMA")
+                       );
   }
-  parse_options_int = (int)NUM2INT(rb_funcall(parse_options, rb_intern("to_i"), 0));
 
-  ctx = xmlSchemaNewMemParserCtxt((const char *)StringValuePtr(content), (int)RSTRING_LEN(content));
-
-  errors = rb_ary_new();
-  xmlSetStructuredErrorFunc((void *)errors, Nokogiri_error_array_pusher);
+  rb_errors = rb_ary_new();
+  xmlSetStructuredErrorFunc((void *)rb_errors, Nokogiri_error_array_pusher);
 
 #ifdef HAVE_XMLSCHEMASETPARSERSTRUCTUREDERRORS
   xmlSchemaSetParserStructuredErrors(
-    ctx,
+    c_parser_context,
     Nokogiri_error_array_pusher,
-    (void *)errors
+    (void *)rb_errors
   );
 #endif
 
+  parse_options_int = (int)NUM2INT(rb_funcall(rb_parse_options, rb_intern("to_i"), 0));
   if (parse_options_int & XML_PARSE_NONET) {
     old_loader = xmlGetExternalEntityLoader();
     xmlSetExternalEntityLoader(xmlNoNetExternalEntityLoader);
   }
 
-  schema = xmlSchemaParse(ctx);
+  c_schema = xmlSchemaParse(c_parser_context);
 
   if (old_loader) {
     xmlSetExternalEntityLoader(old_loader);
   }
 
   xmlSetStructuredErrorFunc(NULL, NULL);
-  xmlSchemaFreeParserCtxt(ctx);
+  xmlSchemaFreeParserCtxt(c_parser_context);
 
-  if (NULL == schema) {
+  if (NULL == c_schema) {
     xmlErrorPtr error = xmlGetLastError();
     if (error) {
       Nokogiri_error_raise(NULL, error);
@@ -155,129 +156,106 @@ read_memory(int argc, VALUE *argv, VALUE klass)
     return Qnil;
   }
 
-  rb_schema = Data_Wrap_Struct(klass, 0, dealloc, schema);
-  rb_iv_set(rb_schema, "@errors", errors);
-  rb_iv_set(rb_schema, "@parse_options", parse_options);
+  rb_schema = TypedData_Wrap_Struct(klass, &xml_schema_type, c_schema);
+  rb_iv_set(rb_schema, "@errors", rb_errors);
+  rb_iv_set(rb_schema, "@parse_options", rb_parse_options);
 
   return rb_schema;
 }
 
-/* Schema creation will remove and deallocate "blank" nodes.
- * If those blank nodes have been exposed to Ruby, they could get freed
- * out from under the VALUE pointer.  This function checks to see if any of
- * those nodes have been exposed to Ruby, and if so we should raise an exception.
+/*
+ * call-seq:
+ *   read_memory(string) → Nokogiri::XML::Schema
+ *
+ * Create a new schema parsed from the contents of +string+
+ *
+ * [Parameters]
+ * - +string+: String containing XML to be parsed as a schema
+ *
+ * [Returns] Nokogiri::XML::Schema
  */
-static int
-has_blank_nodes_p(VALUE cache)
+static VALUE
+read_memory(int argc, VALUE *argv, VALUE klass)
 {
-  long i;
+  VALUE rb_content;
+  VALUE rb_parse_options;
+  xmlSchemaParserCtxtPtr c_parser_context;
 
-  if (NIL_P(cache)) {
-    return 0;
-  }
+  rb_scan_args(argc, argv, "11", &rb_content, &rb_parse_options);
 
-  for (i = 0; i < RARRAY_LEN(cache); i++) {
-    xmlNodePtr node;
-    VALUE element = rb_ary_entry(cache, i);
-    Noko_Node_Get_Struct(element, xmlNode, node);
-    if (xmlIsBlankNode(node)) {
-      return 1;
-    }
-  }
+  c_parser_context = xmlSchemaNewMemParserCtxt(
+                       (const char *)StringValuePtr(rb_content),
+                       (int)RSTRING_LEN(rb_content)
+                     );
 
-  return 0;
+  return xml_schema_parse_schema(klass, c_parser_context, rb_parse_options);
 }
 
 /*
  * call-seq:
- *  from_document(doc)
+ *   from_document(document) → Nokogiri::XML::Schema
  *
- * Create a new Schema from the Nokogiri::XML::Document +doc+
+ * Create a new schema parsed from the +document+.
+ *
+ * [Parameters]
+ * - +document+: Nokogiri::XML::Document to be parsed
+ *
+ * [Returns] Nokogiri::XML::Schema
  */
 static VALUE
-from_document(int argc, VALUE *argv, VALUE klass)
+rb_xml_schema_s_from_document(int argc, VALUE *argv, VALUE klass)
 {
-  VALUE document;
-  VALUE parse_options;
-  int parse_options_int;
-  xmlDocPtr doc;
-  xmlSchemaParserCtxtPtr ctx;
-  xmlSchemaPtr schema;
-  VALUE errors;
+  VALUE rb_document;
+  VALUE rb_parse_options;
   VALUE rb_schema;
-  int scanned_args = 0;
-  xmlExternalEntityLoader old_loader = 0;
+  xmlDocPtr c_document;
+  xmlSchemaParserCtxtPtr c_parser_context;
+  int defensive_copy_p = 0;
 
-  scanned_args = rb_scan_args(argc, argv, "11", &document, &parse_options);
+  rb_scan_args(argc, argv, "11", &rb_document, &rb_parse_options);
 
-  Noko_Node_Get_Struct(document, xmlDoc, doc);
-  doc = doc->doc; /* In case someone passes us a node. ugh. */
-
-  if (scanned_args == 1) {
-    parse_options = rb_const_get_at(rb_const_get_at(mNokogiriXml, rb_intern("ParseOptions")), rb_intern("DEFAULT_SCHEMA"));
-  }
-  parse_options_int = (int)NUM2INT(rb_funcall(parse_options, rb_intern("to_i"), 0));
-
-  if (has_blank_nodes_p(DOC_NODE_CACHE(doc))) {
-    rb_raise(rb_eArgError, "Creating a schema from a document that has blank nodes exposed to Ruby is dangerous");
+  if (!rb_obj_is_kind_of(rb_document, cNokogiriXmlNode)) {
+    rb_raise(rb_eTypeError,
+             "expected parameter to be a Nokogiri::XML::Document, received %"PRIsVALUE,
+             rb_obj_class(rb_document));
   }
 
-  ctx = xmlSchemaNewDocParserCtxt(doc);
-
-  errors = rb_ary_new();
-  xmlSetStructuredErrorFunc((void *)errors, Nokogiri_error_array_pusher);
-
-#ifdef HAVE_XMLSCHEMASETPARSERSTRUCTUREDERRORS
-  xmlSchemaSetParserStructuredErrors(
-    ctx,
-    Nokogiri_error_array_pusher,
-    (void *)errors
-  );
-#endif
-
-  if (parse_options_int & XML_PARSE_NONET) {
-    old_loader = xmlGetExternalEntityLoader();
-    xmlSetExternalEntityLoader(xmlNoNetExternalEntityLoader);
+  if (!rb_obj_is_kind_of(rb_document, cNokogiriXmlDocument)) {
+    xmlNodePtr deprecated_node_type_arg;
+    // TODO: deprecate allowing Node
+    NOKO_WARN_DEPRECATION("Passing a Node as the first parameter to Schema.from_document is deprecated. Please pass a Document instead. This will become an error in a future release of Nokogiri.");
+    Noko_Node_Get_Struct(rb_document, xmlNode, deprecated_node_type_arg);
+    c_document = deprecated_node_type_arg->doc;
+  } else {
+    c_document = noko_xml_document_unwrap(rb_document);
   }
 
-  schema = xmlSchemaParse(ctx);
-
-  if (old_loader) {
-    xmlSetExternalEntityLoader(old_loader);
+  if (noko_xml_document_has_wrapped_blank_nodes_p(c_document)) {
+    // see https://github.com/sparklemotion/nokogiri/pull/2001
+    c_document = xmlCopyDoc(c_document, 1);
+    defensive_copy_p = 1;
   }
 
-  xmlSetStructuredErrorFunc(NULL, NULL);
-  xmlSchemaFreeParserCtxt(ctx);
+  c_parser_context = xmlSchemaNewDocParserCtxt(c_document);
+  rb_schema = xml_schema_parse_schema(klass, c_parser_context, rb_parse_options);
 
-  if (NULL == schema) {
-    xmlErrorPtr error = xmlGetLastError();
-    if (error) {
-      Nokogiri_error_raise(NULL, error);
-    } else {
-      rb_raise(rb_eRuntimeError, "Could not parse document");
-    }
-
-    return Qnil;
+  if (defensive_copy_p) {
+    xmlFreeDoc(c_document);
+    c_document = NULL;
   }
 
-  rb_schema = Data_Wrap_Struct(klass, 0, dealloc, schema);
-  rb_iv_set(rb_schema, "@errors", errors);
-  rb_iv_set(rb_schema, "@parse_options", parse_options);
-
   return rb_schema;
-
-  return Qnil;
 }
 
 void
-noko_init_xml_schema()
+noko_init_xml_schema(void)
 {
   cNokogiriXmlSchema = rb_define_class_under(mNokogiriXml, "Schema", rb_cObject);
 
   rb_undef_alloc_func(cNokogiriXmlSchema);
 
   rb_define_singleton_method(cNokogiriXmlSchema, "read_memory", read_memory, -1);
-  rb_define_singleton_method(cNokogiriXmlSchema, "from_document", from_document, -1);
+  rb_define_singleton_method(cNokogiriXmlSchema, "from_document", rb_xml_schema_s_from_document, -1);
 
   rb_define_private_method(cNokogiriXmlSchema, "validate_document", validate_document, 1);
   rb_define_private_method(cNokogiriXmlSchema, "validate_file",     validate_file, 1);

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

@@ -9,25 +9,39 @@ VALUE cNokogiriXmlText ;
  * Create a new Text element on the +document+ with +content+
  */
 static VALUE
-new (int argc, VALUE *argv, VALUE klass)
+rb_xml_text_s_new(int argc, VALUE *argv, VALUE klass)
 {
-  xmlDocPtr doc;
-  xmlNodePtr node;
-  VALUE string;
-  VALUE document;
-  VALUE rest;
+  xmlDocPtr c_document;
+  xmlNodePtr c_node;
+  VALUE rb_string;
+  VALUE rb_document;
+  VALUE rb_rest;
   VALUE rb_node;
 
-  rb_scan_args(argc, argv, "2*", &string, &document, &rest);
+  rb_scan_args(argc, argv, "2*", &rb_string, &rb_document, &rb_rest);
 
-  Noko_Node_Get_Struct(document, xmlDoc, doc);
+  if (!rb_obj_is_kind_of(rb_document, cNokogiriXmlNode)) {
+    rb_raise(rb_eTypeError,
+             "expected second parameter to be a Nokogiri::XML::Document, received %"PRIsVALUE,
+             rb_obj_class(rb_document));
+  }
 
-  node = xmlNewText((xmlChar *)StringValueCStr(string));
-  node->doc = doc->doc;
+  if (!rb_obj_is_kind_of(rb_document, cNokogiriXmlDocument)) {
+    xmlNodePtr deprecated_node_type_arg;
+    // TODO: deprecate allowing Node
+    NOKO_WARN_DEPRECATION("Passing a Node as the second parameter to Text.new is deprecated. Please pass a Document instead. This will become an error in a future release of Nokogiri.");
+    Noko_Node_Get_Struct(rb_document, xmlNode, deprecated_node_type_arg);
+    c_document = deprecated_node_type_arg->doc;
+  } else {
+    c_document = noko_xml_document_unwrap(rb_document);
+  }
 
-  noko_xml_document_pin_node(node);
+  c_node = xmlNewText((xmlChar *)StringValueCStr(rb_string));
+  c_node->doc = c_document;
 
-  rb_node = noko_xml_node_wrap(klass, node) ;
+  noko_xml_document_pin_node(c_node);
+
+  rb_node = noko_xml_node_wrap(klass, c_node) ;
   rb_obj_call_init(rb_node, argc, argv);
 
   if (rb_block_given_p()) { rb_yield(rb_node); }
@@ -36,7 +50,7 @@ new (int argc, VALUE *argv, VALUE klass)
 }
 
 void
-noko_init_xml_text()
+noko_init_xml_text(void)
 {
   assert(cNokogiriXmlCharacterData);
   /*
@@ -44,5 +58,5 @@ noko_init_xml_text()
    */
   cNokogiriXmlText = rb_define_class_under(mNokogiriXml, "Text", cNokogiriXmlCharacterData);
 
-  rb_define_singleton_method(cNokogiriXmlText, "new", new, -1);
+  rb_define_singleton_method(cNokogiriXmlText, "new", rb_xml_text_s_new, -1);
 }