nokogiri 1.11.0.rc1 → 1.11.1

data/ext/nokogiri/test_global_handlers.c

@@ -0,0 +1,41 @@
+#include <nokogiri.h>
+#include "libxml/xmlerror.h"
+
+static VALUE foreign_error_handler_block = Qnil;
+
+static void foreign_error_handler(void* user_data, xmlErrorPtr c_error)
+{
+  rb_funcall(foreign_error_handler_block, rb_intern("call"), 0);
+}
+
+/*
+ * call-seq:
+ *   __foreign_error_handler { ... } -> nil
+ *
+ * Override libxml2's global error handlers to call the block. This method thus has very little
+ * value except to test that Nokogiri is properly setting error handlers elsewhere in the code. See
+ * test/helper.rb for how this is being used.
+ */
+static VALUE
+rb_foreign_error_handler(VALUE klass)
+{
+  rb_need_block();
+  foreign_error_handler_block = rb_block_proc();
+  xmlSetStructuredErrorFunc(NULL, foreign_error_handler);
+  return Qnil;
+}
+
+/*
+ *  Document-module: Nokogiri::Test
+ *
+ *  The Nokogiri::Test module should only be used for testing Nokogiri.
+ *  Do NOT use this outside of the Nokogiri test suite.
+ */
+void
+init_test_global_handlers()
+{
+  VALUE mNokogiri = rb_define_module("Nokogiri");
+  VALUE mNokogiriTest = rb_define_module_under(mNokogiri, "Test");
+
+  rb_define_singleton_method(mNokogiriTest, "__foreign_error_handler", rb_foreign_error_handler, 0);
+}

data/ext/nokogiri/xml_document.c

@@ -1,6 +1,6 @@
 #include <xml_document.h>
 
-static int dealloc_node_i(xmlNodePtr key, xmlNodePtr node, xmlDocPtr doc)
+static int dealloc_node_i2(xmlNodePtr key, xmlNodePtr node, xmlDocPtr doc)
 {
   switch(node->type) {
   case XML_ATTRIBUTE_NODE:
@@ -20,6 +20,11 @@ static int dealloc_node_i(xmlNodePtr key, xmlNodePtr node, xmlDocPtr doc)
   return ST_CONTINUE;
 }
 
+static int dealloc_node_i(st_data_t key, st_data_t node, st_data_t doc)
+{
+  return dealloc_node_i2((xmlNodePtr)key, (xmlNodePtr)node, (xmlDocPtr)doc);
+}
+
 static void remove_private(xmlNodePtr node)
 {
   xmlNodePtr child;
@@ -501,7 +506,7 @@ static int block_caller(void * ctx, xmlNodePtr _node, xmlNodePtr _parent)
  * The block must return a non-nil, non-false value if the +obj+ passed in
  * should be included in the canonicalized document.
  */
-static VALUE canonicalize(int argc, VALUE* argv, VALUE self)
+static VALUE nokogiri_xml_document_canonicalize(int argc, VALUE* argv, VALUE self)
 {
   VALUE mode;
   VALUE incl_ns;
@@ -582,7 +587,7 @@ void init_xml_document()
   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, "canonicalize", canonicalize, -1);
+  rb_define_method(klass, "canonicalize", nokogiri_xml_document_canonicalize, -1);
   rb_define_method(klass, "dup", duplicate_document, -1);
   rb_define_method(klass, "url", url, 0);
   rb_define_method(klass, "create_entity", create_entity, -1);

data/ext/nokogiri/xml_io.c

@@ -2,12 +2,13 @@
 
 static ID id_read, id_write;
 
-VALUE read_check(VALUE *args) {
+VALUE read_check(VALUE val) {
+  VALUE *args = (VALUE *)val;
   return rb_funcall(args[0], id_read, 1, args[1]);
 }
 
-VALUE read_failed(void) {
-	return Qundef;
+VALUE read_failed(VALUE arg, VALUE exc) {
+  return Qundef;
 }
 
 int io_read_callback(void * ctx, char * buffer, int len) {
@@ -30,12 +31,13 @@ int io_read_callback(void * ctx, char * buffer, int len) {
   return (int)safe_len;
 }
 
-VALUE write_check(VALUE *args) {
+VALUE write_check(VALUE val) {
+  VALUE *args = (VALUE *)val;
   return rb_funcall(args[0], id_write, 1, args[1]);
 }
 
-VALUE write_failed(void) {
-	return Qundef;
+VALUE write_failed(VALUE arg, VALUE exc) {
+  return Qundef;
 }
 
 int io_write_callback(void * ctx, char * buffer, int len) {

data/ext/nokogiri/xml_node.c

@@ -301,7 +301,7 @@ ok:
        *  issue #391, where new node's prefix may become the string "default"
        *  see libxml2 tree.c xmlNewReconciliedNs which implements this behavior.
        */
-      xmlFree(reparentee->ns->prefix);
+      xmlFree((xmlChar*)reparentee->ns->prefix);
       reparentee->ns->prefix = NULL;
     }
   }

data/ext/nokogiri/xml_node_set.c

@@ -319,7 +319,7 @@ static VALUE slice(int argc, VALUE *argv, VALUE self)
  *
  * Return this list as an Array
  */
-static VALUE to_array(VALUE self, VALUE rb_node)
+static VALUE to_array(VALUE self)
 {
   xmlNodeSetPtr node_set ;
   VALUE list;

data/ext/nokogiri/xml_reader.c

@@ -28,35 +28,24 @@ static int has_attributes(xmlTextReaderPtr reader)
 static void Nokogiri_xml_node_namespaces(xmlNodePtr node, VALUE attr_hash)
 {
   xmlNsPtr ns;
-  static char buffer[XMLNS_BUFFER_LEN] ;
-  char *key ;
-  size_t keylen ;
+  VALUE key;
 
   if (node->type != XML_ELEMENT_NODE) return ;
 
   ns = node->nsDef;
   while (ns != NULL) {
 
-    keylen = XMLNS_PREFIX_LEN + (ns->prefix ? (strlen((const char*)ns->prefix) + 1) : 0) ;
-    if (keylen > XMLNS_BUFFER_LEN) {
-      key = (char*)malloc(keylen) ;
-    } else {
-      key = buffer ;
-    }
-
+    key = rb_enc_str_new_cstr(XMLNS_PREFIX, rb_utf8_encoding());
     if (ns->prefix) {
-      sprintf(key, "%s:%s", XMLNS_PREFIX, ns->prefix);
-    } else {
-      sprintf(key, "%s", XMLNS_PREFIX);
+      rb_str_cat_cstr(key, ":");
+      rb_str_cat_cstr(key, (const char*)ns->prefix);
     }
 
+    key = rb_str_conv_enc(key, rb_utf8_encoding(), rb_default_internal_encoding());
     rb_hash_aset(attr_hash,
-        NOKOGIRI_STR_NEW2(key),
+        key,
         (ns->href ? NOKOGIRI_STR_NEW2(ns->href) : Qnil)
     );
-    if (key != buffer) {
-      free(key);
-    }
     ns = ns->next ;
   }
 }

data/ext/nokogiri/xml_relax_ng.c

@@ -53,16 +53,24 @@ static VALUE validate_document(VALUE self, VALUE document)
  *
  * Create a new RelaxNG from the contents of +string+
  */
-static VALUE read_memory(VALUE klass, VALUE content)
+static VALUE read_memory(int argc, VALUE *argv, VALUE klass)
 {
-  xmlRelaxNGParserCtxtPtr ctx = xmlRelaxNGNewMemParserCtxt(
-      (const char *)StringValuePtr(content),
-      (int)RSTRING_LEN(content)
-  );
+  VALUE content;
+  VALUE parse_options;
+  xmlRelaxNGParserCtxtPtr ctx;
   xmlRelaxNGPtr schema;
-  VALUE errors = rb_ary_new();
+  VALUE errors;
   VALUE rb_schema;
+  int scanned_args = 0;
+
+  scanned_args = rb_scan_args(argc, argv, "11", &content, &parse_options);
+  if (scanned_args == 1) {
+    parse_options = rb_const_get(rb_const_get(mNokogiriXml, rb_intern("ParseOptions")), rb_intern("DEFAULT_SCHEMA"));
+  }
 
+  ctx = xmlRelaxNGNewMemParserCtxt((const char *)StringValuePtr(content), (int)RSTRING_LEN(content));
+
+  errors = rb_ary_new();
   xmlSetStructuredErrorFunc((void *)errors, Nokogiri_error_array_pusher);
 
 #ifdef HAVE_XMLRELAXNGSETPARSERSTRUCTUREDERRORS
@@ -90,6 +98,7 @@ static VALUE read_memory(VALUE klass, VALUE content)
 
   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;
 }
@@ -100,18 +109,25 @@ static VALUE read_memory(VALUE klass, VALUE content)
  *
  * Create a new RelaxNG schema from the Nokogiri::XML::Document +doc+
  */
-static VALUE from_document(VALUE klass, VALUE document)
+static VALUE from_document(int argc, VALUE *argv, VALUE klass)
 {
+  VALUE document;
+  VALUE parse_options;
   xmlDocPtr doc;
   xmlRelaxNGParserCtxtPtr ctx;
   xmlRelaxNGPtr schema;
   VALUE errors;
   VALUE rb_schema;
+  int scanned_args = 0;
+
+  scanned_args = rb_scan_args(argc, argv, "11", &document, &parse_options);
 
   Data_Get_Struct(document, xmlDoc, doc);
+  doc = doc->doc; /* In case someone passes us a node. ugh. */
 
-  /* In case someone passes us a node. ugh. */
-  doc = doc->doc;
+  if (scanned_args == 1) {
+    parse_options = rb_const_get(rb_const_get(mNokogiriXml, rb_intern("ParseOptions")), rb_intern("DEFAULT_SCHEMA"));
+  }
 
   ctx = xmlRelaxNGNewDocParserCtxt(doc);
 
@@ -129,6 +145,7 @@ static VALUE from_document(VALUE klass, VALUE document)
   schema = xmlRelaxNGParse(ctx);
 
   xmlSetStructuredErrorFunc(NULL, NULL);
+  xmlRelaxNGFreeParserCtxt(ctx);
 
   if(NULL == schema) {
     xmlErrorPtr error = xmlGetLastError();
@@ -142,6 +159,7 @@ static VALUE from_document(VALUE klass, VALUE document)
 
   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;
 }
@@ -155,7 +173,7 @@ void init_xml_relax_ng()
 
   cNokogiriXmlRelaxNG = klass;
 
-  rb_define_singleton_method(klass, "read_memory", read_memory, 1);
-  rb_define_singleton_method(klass, "from_document", from_document, 1);
+  rb_define_singleton_method(klass, "read_memory", read_memory, -1);
+  rb_define_singleton_method(klass, "from_document", from_document, -1);
   rb_define_private_method(klass, "validate_document", validate_document, 1);
 }

data/ext/nokogiri/xml_sax_parser.c

@@ -1,8 +1,5 @@
 #include <xml_sax_parser.h>
 
-int vasprintf (char **strp, const char *fmt, va_list ap);
-void vasprintf_free (void *p);
-
 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;
@@ -206,7 +203,7 @@ static void warning_func(void * ctx, const char *msg, ...)
   va_end(args);
 
   ruby_message = NOKOGIRI_STR_NEW2(message);
-  vasprintf_free(message);
+  free(message);
   rb_funcall(doc, id_warning, 1, ruby_message);
 }
 
@@ -223,7 +220,7 @@ static void error_func(void * ctx, const char *msg, ...)
   va_end(args);
 
   ruby_message = NOKOGIRI_STR_NEW2(message);
-  vasprintf_free(message);
+  free(message);
   rb_funcall(doc, id_error, 1, ruby_message);
 }
 
@@ -262,8 +259,6 @@ static VALUE allocate(VALUE klass)
 {
   xmlSAXHandlerPtr handler = calloc((size_t)1, sizeof(xmlSAXHandler));
 
-  xmlSetStructuredErrorFunc(NULL, NULL);
-
   handler->startDocument = start_document;
   handler->endDocument = end_document;
   handler->startElement = start_element;

data/ext/nokogiri/xml_sax_parser_context.c

@@ -4,13 +4,13 @@ VALUE cNokogiriXmlSaxParserContext ;
 
 static void deallocate(xmlParserCtxtPtr ctxt)
 {
-  NOKOGIRI_DEBUG_START(handler);
+  NOKOGIRI_DEBUG_START(ctxt);
 
   ctxt->sax = NULL;
 
   xmlFreeParserCtxt(ctxt);
 
-  NOKOGIRI_DEBUG_END(handler);
+  NOKOGIRI_DEBUG_END(ctxt);
 }
 
 /*
@@ -120,6 +120,8 @@ parse_with(VALUE self, VALUE sax_handler)
     ctxt->sax = sax;
     ctxt->userData = (void *)NOKOGIRI_SAX_TUPLE_NEW(ctxt, sax_handler);
 
+    xmlSetStructuredErrorFunc(NULL, NULL);
+
     rb_ensure(parse_doc, (VALUE)ctxt, parse_doc_finalize, (VALUE)ctxt);
 
     return Qnil;

data/ext/nokogiri/xml_sax_push_parser.c

@@ -35,6 +35,8 @@ static VALUE native_write(VALUE self, VALUE _chunk, VALUE _last_chunk)
     size = (int)RSTRING_LEN(_chunk);
   }
 
+  xmlSetStructuredErrorFunc(NULL, NULL);
+
   if (xmlParseChunk(ctx, chunk, size, Qtrue == _last_chunk ? 1 : 0)) {
     if (!(ctx->options & XML_PARSE_RECOVER)) {
       xmlErrorPtr e = xmlCtxtGetLastError(ctx);

data/ext/nokogiri/xml_schema.c

@@ -93,15 +93,27 @@ static VALUE validate_file(VALUE self, VALUE rb_filename)
  *
  * Create a new Schema from the contents of +string+
  */
-static VALUE read_memory(VALUE klass, VALUE content)
+static VALUE read_memory(int argc, VALUE *argv, VALUE klass)
 {
+  VALUE content;
+  VALUE parse_options;
+  int parse_options_int;
+  xmlSchemaParserCtxtPtr ctx;
   xmlSchemaPtr schema;
-  xmlSchemaParserCtxtPtr ctx = xmlSchemaNewMemParserCtxt(
-      (const char *)StringValuePtr(content),
-      (int)RSTRING_LEN(content)
-  );
+  VALUE errors;
   VALUE rb_schema;
-  VALUE errors = rb_ary_new();
+  int scanned_args = 0;
+  xmlExternalEntityLoader old_loader = 0;
+
+  scanned_args = rb_scan_args(argc, argv, "11", &content, &parse_options);
+  if (scanned_args == 1) {
+    parse_options = rb_const_get(rb_const_get(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);
 
 #ifdef HAVE_XMLSCHEMASETPARSERSTRUCTUREDERRORS
@@ -109,10 +121,19 @@ static VALUE read_memory(VALUE klass, VALUE content)
     ctx,
     Nokogiri_error_array_pusher,
     (void *)errors
-  );
+    );
 #endif
 
-   schema = xmlSchemaParse(ctx);
+  if (parse_options_int & XML_PARSE_NONET) {
+    old_loader = xmlGetExternalEntityLoader();
+    xmlSetExternalEntityLoader(xmlNoNetExternalEntityLoader);
+  }
+
+  schema = xmlSchemaParse(ctx);
+
+  if (old_loader) {
+    xmlSetExternalEntityLoader(old_loader);
+  }
 
   xmlSetStructuredErrorFunc(NULL, NULL);
   xmlSchemaFreeParserCtxt(ctx);
@@ -129,28 +150,68 @@ static VALUE read_memory(VALUE klass, VALUE content)
 
   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;
 }
 
+/* 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.
+ */
+static int has_blank_nodes_p(VALUE cache)
+{
+    long i;
+
+    if (NIL_P(cache)) {
+        return 0;
+    }
+
+    for (i = 0; i < RARRAY_LEN(cache); i++) {
+        xmlNodePtr node;
+        VALUE element = rb_ary_entry(cache, i);
+        Data_Get_Struct(element, xmlNode, node);
+        if (xmlIsBlankNode(node)) {
+            return 1;
+        }
+    }
+
+    return 0;
+}
+
 /*
  * call-seq:
  *  from_document(doc)
  *
  * Create a new Schema from the Nokogiri::XML::Document +doc+
  */
-static VALUE from_document(VALUE klass, VALUE document)
+static VALUE 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_schema;
+  int scanned_args = 0;
+  xmlExternalEntityLoader old_loader = 0;
+
+  scanned_args = rb_scan_args(argc, argv, "11", &document, &parse_options);
 
   Data_Get_Struct(document, xmlDoc, doc);
+  doc = doc->doc; /* In case someone passes us a node. ugh. */
 
-  /* In case someone passes us a node. ugh. */
-  doc = doc->doc;
+  if (scanned_args == 1) {
+    parse_options = rb_const_get(rb_const_get(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");
+  }
 
   ctx = xmlSchemaNewDocParserCtxt(doc);
 
@@ -165,8 +226,17 @@ static VALUE from_document(VALUE klass, VALUE document)
   );
 #endif
 
+  if (parse_options_int & XML_PARSE_NONET) {
+    old_loader = xmlGetExternalEntityLoader();
+    xmlSetExternalEntityLoader(xmlNoNetExternalEntityLoader);
+  }
+
   schema = xmlSchemaParse(ctx);
 
+  if (old_loader) {
+    xmlSetExternalEntityLoader(old_loader);
+  }
+
   xmlSetStructuredErrorFunc(NULL, NULL);
   xmlSchemaFreeParserCtxt(ctx);
 
@@ -182,6 +252,7 @@ static VALUE from_document(VALUE klass, VALUE document)
 
   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;
 
@@ -197,8 +268,8 @@ void init_xml_schema()
 
   cNokogiriXmlSchema = klass;
 
-  rb_define_singleton_method(klass, "read_memory", read_memory, 1);
-  rb_define_singleton_method(klass, "from_document", from_document, 1);
+  rb_define_singleton_method(klass, "read_memory", read_memory, -1);
+  rb_define_singleton_method(klass, "from_document", from_document, -1);
 
   rb_define_private_method(klass, "validate_document", validate_document, 1);
   rb_define_private_method(klass, "validate_file",     validate_file, 1);