nokogiri 1.6.0 → 1.13.2

data/ext/nokogiri/xml_relax_ng.c

@@ -1,6 +1,9 @@
-#include <xml_relax_ng.h>
+#include <nokogiri.h>
 
-static void dealloc(xmlRelaxNGPtr schema)
+VALUE cNokogiriXmlRelaxNG;
+
+static void
+dealloc(xmlRelaxNGPtr schema)
 {
   NOKOGIRI_DEBUG_START(schema);
   xmlRelaxNGFree(schema);
@@ -13,7 +16,8 @@ static void dealloc(xmlRelaxNGPtr schema)
  *
  * Validate a Nokogiri::XML::Document against this RelaxNG schema.
  */
-static VALUE validate_document(VALUE self, VALUE document)
+static VALUE
+validate_document(VALUE self, VALUE document)
 {
   xmlDocPtr doc;
   xmlRelaxNGPtr schema;
@@ -27,7 +31,7 @@ static VALUE validate_document(VALUE self, VALUE document)
 
   valid_ctxt = xmlRelaxNGNewValidCtxt(schema);
 
-  if(NULL == valid_ctxt) {
+  if (NULL == valid_ctxt) {
     /* we have a problem */
     rb_raise(rb_eRuntimeError, "Could not create a validation context");
   }
@@ -53,16 +57,25 @@ 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_at(rb_const_get_at(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
@@ -78,18 +91,20 @@ static VALUE read_memory(VALUE klass, VALUE content)
   xmlSetStructuredErrorFunc(NULL, NULL);
   xmlRelaxNGFreeParserCtxt(ctx);
 
-  if(NULL == schema) {
+  if (NULL == schema) {
     xmlErrorPtr error = xmlGetLastError();
-    if(error)
+    if (error) {
       Nokogiri_error_raise(NULL, error);
-    else
+    } else {
       rb_raise(rb_eRuntimeError, "Could not parse document");
+    }
 
     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);
 
   return rb_schema;
 }
@@ -100,18 +115,26 @@ 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_at(rb_const_get_at(mNokogiriXml, rb_intern("ParseOptions")), rb_intern("DEFAULT_SCHEMA"));
+  }
 
   ctx = xmlRelaxNGNewDocParserCtxt(doc);
 
@@ -129,33 +152,34 @@ static VALUE from_document(VALUE klass, VALUE document)
   schema = xmlRelaxNGParse(ctx);
 
   xmlSetStructuredErrorFunc(NULL, NULL);
+  xmlRelaxNGFreeParserCtxt(ctx);
 
-  if(NULL == schema) {
+  if (NULL == schema) {
     xmlErrorPtr error = xmlGetLastError();
-    if(error)
+    if (error) {
       Nokogiri_error_raise(NULL, error);
-    else
+    } else {
       rb_raise(rb_eRuntimeError, "Could not parse document");
+    }
 
     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);
 
   return rb_schema;
 }
 
-VALUE cNokogiriXmlRelaxNG;
-void init_xml_relax_ng()
+void
+noko_init_xml_relax_ng()
 {
-  VALUE nokogiri = rb_define_module("Nokogiri");
-  VALUE xml = rb_define_module_under(nokogiri, "XML");
-  VALUE klass = rb_define_class_under(xml, "RelaxNG", cNokogiriXmlSchema);
+  assert(cNokogiriXmlSchema);
+  cNokogiriXmlRelaxNG = rb_define_class_under(mNokogiriXml, "RelaxNG", cNokogiriXmlSchema);
 
-  cNokogiriXmlRelaxNG = klass;
+  rb_define_singleton_method(cNokogiriXmlRelaxNG, "read_memory", read_memory, -1);
+  rb_define_singleton_method(cNokogiriXmlRelaxNG, "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(cNokogiriXmlRelaxNG, "validate_document", validate_document, 1);
 }

data/ext/nokogiri/xml_sax_parser.c

@@ -1,38 +1,35 @@
-#include <xml_sax_parser.h>
+#include <nokogiri.h>
 
-int vasprintf (char **strp, const char *fmt, va_list ap);
-void vasprintf_free (void *p);
+VALUE cNokogiriXmlSaxParser ;
 
 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;
-static ID id_cdata_block, id_cAttribute;
+static ID id_cdata_block;
 static ID id_processing_instruction;
 
-#define STRING_OR_NULL(str) \
-   (RTEST(str) ? StringValuePtr(str) : NULL)
-
-static void start_document(void * ctx)
+static void
+start_document(void *ctx)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
 
   xmlParserCtxtPtr ctxt = NOKOGIRI_SAX_CTXT(ctx);
 
-  if(NULL != ctxt && ctxt->html != 1) {
-    if(ctxt->standalone != -1) {  /* -1 means there was no declaration */
-      VALUE encoding = ctxt->encoding ?
-        NOKOGIRI_STR_NEW2(ctxt->encoding) :
-        Qnil;
-
-      VALUE version = ctxt->version ?
-        NOKOGIRI_STR_NEW2(ctxt->version) :
-        Qnil;
-
+  if (NULL != ctxt && ctxt->html != 1) {
+    if (ctxt->standalone != -1) { /* -1 means there was no declaration */
+      VALUE encoding = Qnil ;
       VALUE standalone = Qnil;
+      VALUE version;
+      if (ctxt->encoding) {
+        encoding = NOKOGIRI_STR_NEW2(ctxt->encoding) ;
+      } else if (ctxt->input && ctxt->input->encoding) {
+        encoding = NOKOGIRI_STR_NEW2(ctxt->input->encoding) ;
+      }
 
-      switch(ctxt->standalone)
-      {
+      version = ctxt->version ? NOKOGIRI_STR_NEW2(ctxt->version) : Qnil;
+
+      switch (ctxt->standalone) {
         case 0:
           standalone = NOKOGIRI_STR_NEW2("no");
           break;
@@ -48,138 +45,140 @@ static void start_document(void * ctx)
   rb_funcall(doc, id_start_document, 0);
 }
 
-static void end_document(void * ctx)
+static void
+end_document(void *ctx)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
   rb_funcall(doc, id_end_document, 0);
 }
 
-static void start_element(void * ctx, const xmlChar *name, const xmlChar **atts)
+static void
+start_element(void *ctx, const xmlChar *name, const xmlChar **atts)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
   VALUE attributes = rb_ary_new();
-  const xmlChar * attr;
+  const xmlChar *attr;
   int i = 0;
-  if(atts) {
-    while((attr = atts[i]) != NULL) {
-      const xmlChar * val = atts[i+1];
+  if (atts) {
+    while ((attr = atts[i]) != NULL) {
+      const xmlChar *val = atts[i + 1];
       VALUE value = val != NULL ? NOKOGIRI_STR_NEW2(val) : Qnil;
       rb_ary_push(attributes, rb_ary_new3(2, NOKOGIRI_STR_NEW2(attr), value));
-      i+=2;
+      i += 2;
     }
   }
 
-  rb_funcall( doc,
-              id_start_element,
-              2,
-              NOKOGIRI_STR_NEW2(name),
-              attributes
-  );
+  rb_funcall(doc,
+             id_start_element,
+             2,
+             NOKOGIRI_STR_NEW2(name),
+             attributes
+            );
 }
 
-static void end_element(void * ctx, const xmlChar *name)
+static void
+end_element(void *ctx, const xmlChar *name)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
   rb_funcall(doc, id_end_element, 1, NOKOGIRI_STR_NEW2(name));
 }
 
-static VALUE attributes_as_list(
-  VALUE self,
-  int nb_attributes,
-  const xmlChar ** attributes)
+static VALUE
+attributes_as_array(int attributes_len, const xmlChar **c_attributes)
 {
-  VALUE list = rb_ary_new2((long)nb_attributes);
+  VALUE rb_array = rb_ary_new2((long)attributes_len);
+  VALUE cNokogiriXmlSaxParserAttribute;
 
-  VALUE attr_klass = rb_const_get(cNokogiriXmlSaxParser, id_cAttribute);
-  if (attributes) {
+  cNokogiriXmlSaxParserAttribute = rb_const_get_at(cNokogiriXmlSaxParser, rb_intern("Attribute"));
+  if (c_attributes) {
     /* Each attribute is an array of [localname, prefix, URI, value, end] */
     int i;
-    for (i = 0; i < nb_attributes * 5; i += 5) {
-      VALUE argv[4], attribute;
+    for (i = 0; i < attributes_len * 5; i += 5) {
+      VALUE rb_constructor_args[4], rb_attribute;
 
-      argv[0] = RBSTR_OR_QNIL(attributes[i + 0]); /* localname */
-      argv[1] = RBSTR_OR_QNIL(attributes[i + 1]); /* prefix */
-      argv[2] = RBSTR_OR_QNIL(attributes[i + 2]); /* URI */
+      rb_constructor_args[0] = RBSTR_OR_QNIL(c_attributes[i + 0]); /* localname */
+      rb_constructor_args[1] = RBSTR_OR_QNIL(c_attributes[i + 1]); /* prefix */
+      rb_constructor_args[2] = RBSTR_OR_QNIL(c_attributes[i + 2]); /* URI */
 
       /* value */
-      argv[3] = NOKOGIRI_STR_NEW((const char*)attributes[i+3],
-          (attributes[i+4] - attributes[i+3]));
+      rb_constructor_args[3] = NOKOGIRI_STR_NEW((const char *)c_attributes[i + 3],
+                               (c_attributes[i + 4] - c_attributes[i + 3]));
 
-      attribute = rb_class_new_instance(4, argv, attr_klass);
-      rb_ary_push(list, attribute);
+      rb_attribute = rb_class_new_instance(4, rb_constructor_args, cNokogiriXmlSaxParserAttribute);
+      rb_ary_push(rb_array, rb_attribute);
     }
   }
 
-  return list;
+  return rb_array;
 }
 
 static void
-start_element_ns (
-  void * ctx,
-  const xmlChar * localname,
-  const xmlChar * prefix,
-  const xmlChar * uri,
+start_element_ns(
+  void *ctx,
+  const xmlChar *localname,
+  const xmlChar *prefix,
+  const xmlChar *uri,
   int nb_namespaces,
-  const xmlChar ** namespaces,
+  const xmlChar **namespaces,
   int nb_attributes,
   int nb_defaulted,
-  const xmlChar ** attributes)
+  const xmlChar **attributes)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
 
-  VALUE attribute_list = attributes_as_list(self, nb_attributes, attributes);
+  VALUE attribute_ary = attributes_as_array(nb_attributes, attributes);
 
   VALUE ns_list = rb_ary_new2((long)nb_namespaces);
 
   if (namespaces) {
     int i;
-    for (i = 0; i < nb_namespaces * 2; i += 2)
-    {
+    for (i = 0; i < nb_namespaces * 2; i += 2) {
       rb_ary_push(ns_list,
-        rb_ary_new3((long)2,
-          RBSTR_OR_QNIL(namespaces[i + 0]),
-          RBSTR_OR_QNIL(namespaces[i + 1])
-        )
-      );
+                  rb_ary_new3((long)2,
+                              RBSTR_OR_QNIL(namespaces[i + 0]),
+                              RBSTR_OR_QNIL(namespaces[i + 1])
+                             )
+                 );
     }
   }
 
-  rb_funcall( doc,
-              id_start_element_namespace,
-              5,
-              NOKOGIRI_STR_NEW2(localname),
-              attribute_list,
-              RBSTR_OR_QNIL(prefix),
-              RBSTR_OR_QNIL(uri),
-              ns_list
-  );
+  rb_funcall(doc,
+             id_start_element_namespace,
+             5,
+             NOKOGIRI_STR_NEW2(localname),
+             attribute_ary,
+             RBSTR_OR_QNIL(prefix),
+             RBSTR_OR_QNIL(uri),
+             ns_list
+            );
 }
 
 /**
- * end_element_ns was borrowed heavily from libxml-ruby. 
+ * end_element_ns was borrowed heavily from libxml-ruby.
  */
 static void
-end_element_ns (
-  void * ctx,
-  const xmlChar * localname,
-  const xmlChar * prefix,
-  const xmlChar * uri)
+end_element_ns(
+  void *ctx,
+  const xmlChar *localname,
+  const xmlChar *prefix,
+  const xmlChar *uri)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
 
-  rb_funcall(doc, id_end_element_namespace, 3, 
-    NOKOGIRI_STR_NEW2(localname),
-    RBSTR_OR_QNIL(prefix),
-    RBSTR_OR_QNIL(uri)
-  );
+  rb_funcall(doc, id_end_element_namespace, 3,
+             NOKOGIRI_STR_NEW2(localname),
+             RBSTR_OR_QNIL(prefix),
+             RBSTR_OR_QNIL(uri)
+            );
 }
 
-static void characters_func(void * ctx, const xmlChar * ch, int len)
+static void
+characters_func(void *ctx, const xmlChar *ch, int len)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
@@ -187,7 +186,8 @@ static void characters_func(void * ctx, const xmlChar * ch, int len)
   rb_funcall(doc, id_characters, 1, str);
 }
 
-static void comment_func(void * ctx, const xmlChar * value)
+static void
+comment_func(void *ctx, const xmlChar *value)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
@@ -195,11 +195,12 @@ static void comment_func(void * ctx, const xmlChar * value)
   rb_funcall(doc, id_comment, 1, str);
 }
 
-static void warning_func(void * ctx, const char *msg, ...)
+static void
+warning_func(void *ctx, const char *msg, ...)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
-  char * message;
+  char *message;
   VALUE ruby_message;
 
   va_list args;
@@ -208,15 +209,16 @@ 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);
 }
 
-static void error_func(void * ctx, const char *msg, ...)
+static void
+error_func(void *ctx, const char *msg, ...)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
-  char * message;
+  char *message;
   VALUE ruby_message;
 
   va_list args;
@@ -225,11 +227,12 @@ 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);
 }
 
-static void cdata_block(void * ctx, const xmlChar * value, int len)
+static void
+cdata_block(void *ctx, const xmlChar *value, int len)
 {
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
   VALUE doc = rb_iv_get(self, "@document");
@@ -237,7 +240,8 @@ static void cdata_block(void * ctx, const xmlChar * value, int len)
   rb_funcall(doc, id_cdata_block, 1, string);
 }
 
-static void processing_instruction(void * ctx, const xmlChar * name, const xmlChar * content)
+static void
+processing_instruction(void *ctx, const xmlChar *name, const xmlChar *content)
 {
   VALUE rb_content;
   VALUE self = NOKOGIRI_SAX_SELF(ctx);
@@ -245,27 +249,27 @@ static void processing_instruction(void * ctx, const xmlChar * name, const xmlCh
 
   rb_content = content ? NOKOGIRI_STR_NEW2(content) : Qnil;
 
-  rb_funcall( doc,
-              id_processing_instruction,
-              2,
-              NOKOGIRI_STR_NEW2(name),
-              rb_content
-  );
+  rb_funcall(doc,
+             id_processing_instruction,
+             2,
+             NOKOGIRI_STR_NEW2(name),
+             rb_content
+            );
 }
 
-static void deallocate(xmlSAXHandlerPtr handler)
+static void
+deallocate(xmlSAXHandlerPtr handler)
 {
   NOKOGIRI_DEBUG_START(handler);
   free(handler);
   NOKOGIRI_DEBUG_END(handler);
 }
 
-static VALUE allocate(VALUE klass)
+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;
@@ -283,29 +287,23 @@ static VALUE allocate(VALUE klass)
   return Data_Wrap_Struct(klass, NULL, deallocate, handler);
 }
 
-VALUE cNokogiriXmlSaxParser ;
-void init_xml_sax_parser()
+void
+noko_init_xml_sax_parser()
 {
-  VALUE nokogiri  = rb_define_module("Nokogiri");
-  VALUE xml       = rb_define_module_under(nokogiri, "XML");
-  VALUE sax       = rb_define_module_under(xml, "SAX");
-  VALUE klass     = rb_define_class_under(sax, "Parser", rb_cObject);
-
-  cNokogiriXmlSaxParser = klass;
+  cNokogiriXmlSaxParser = rb_define_class_under(mNokogiriXmlSax, "Parser", rb_cObject);
 
-  rb_define_alloc_func(klass, allocate);
+  rb_define_alloc_func(cNokogiriXmlSaxParser, allocate);
 
   id_start_document = rb_intern("start_document");
-  id_end_document   = rb_intern("end_document");
-  id_start_element  = rb_intern("start_element");
-  id_end_element    = rb_intern("end_element");
-  id_comment        = rb_intern("comment");
-  id_characters     = rb_intern("characters");
-  id_xmldecl        = rb_intern("xmldecl");
-  id_error          = rb_intern("error");
-  id_warning        = rb_intern("warning");
-  id_cdata_block    = rb_intern("cdata_block");
-  id_cAttribute     = rb_intern("Attribute");
+  id_end_document = rb_intern("end_document");
+  id_start_element = rb_intern("start_element");
+  id_end_element = rb_intern("end_element");
+  id_comment = rb_intern("comment");
+  id_characters = rb_intern("characters");
+  id_xmldecl = rb_intern("xmldecl");
+  id_error = rb_intern("error");
+  id_warning = rb_intern("warning");
+  id_cdata_block = rb_intern("cdata_block");
   id_start_element_namespace = rb_intern("start_element_namespace");
   id_end_element_namespace = rb_intern("end_element_namespace");
   id_processing_instruction = rb_intern("processing_instruction");