adding relaxng and xml schema support. GH sparklemotion#4

CHANGELOG.rdoc

@@ -12,6 +12,8 @@
   * Nokogiri::HTML::Document#meta_encoding for getting the meta encoding
   * Nokogiri::HTML::Document#meta_encoding= for setting the meta encoding
   * Nokogiri::XML::Document#encoding= to set the document encoding
+  * Nokogiri::XML::Schema for validating documents against XSD schema
+  * Nokogiri::XML::RelaxNG for validating documents against RelaxNG schema
 
 * Bugfixes
 

Manifest.txt

@@ -38,10 +38,14 @@ ext/nokogiri/xml_processing_instruction.c
 ext/nokogiri/xml_processing_instruction.h
 ext/nokogiri/xml_reader.c
 ext/nokogiri/xml_reader.h
+ext/nokogiri/xml_relax_ng.c
+ext/nokogiri/xml_relax_ng.h
 ext/nokogiri/xml_sax_parser.c
 ext/nokogiri/xml_sax_parser.h
 ext/nokogiri/xml_sax_push_parser.c
 ext/nokogiri/xml_sax_push_parser.h
+ext/nokogiri/xml_schema.c
+ext/nokogiri/xml_schema.h
 ext/nokogiri/xml_syntax_error.c
 ext/nokogiri/xml_syntax_error.h
 ext/nokogiri/xml_text.c
@@ -95,10 +99,12 @@ lib/nokogiri/xml/node_set.rb
 lib/nokogiri/xml/notation.rb
 lib/nokogiri/xml/processing_instruction.rb
 lib/nokogiri/xml/reader.rb
+lib/nokogiri/xml/relax_ng.rb
 lib/nokogiri/xml/sax.rb
 lib/nokogiri/xml/sax/document.rb
 lib/nokogiri/xml/sax/parser.rb
 lib/nokogiri/xml/sax/push_parser.rb
+lib/nokogiri/xml/schema.rb
 lib/nokogiri/xml/syntax_error.rb
 lib/nokogiri/xml/text.rb
 lib/nokogiri/xml/xpath.rb
@@ -112,9 +118,13 @@ test/css/test_nthiness.rb
 test/css/test_parser.rb
 test/css/test_tokenizer.rb
 test/css/test_xpath_visitor.rb
+test/files/address_book.rlx
+test/files/address_book.xml
 test/files/dont_hurt_em_why.xml
 test/files/exslt.xml
 test/files/exslt.xslt
+test/files/po.xml
+test/files/po.xsd
 test/files/staff.xml
 test/files/staff.xslt
 test/files/tlm.html
@@ -167,6 +177,8 @@ test/xml/test_node_encoding.rb
 test/xml/test_node_set.rb
 test/xml/test_processing_instruction.rb
 test/xml/test_reader_encoding.rb
+test/xml/test_relax_ng.rb
+test/xml/test_schema.rb
 test/xml/test_text.rb
 test/xml/test_unparented_node.rb
 test/xml/test_xpath.rb

ext/nokogiri/native.c

@@ -70,4 +70,6 @@ void Init_native()
   init_xslt_stylesheet();
   init_xml_syntax_error();
   init_html_entity_lookup();
+  init_xml_schema();
+  init_xml_relax_ng();
 }

ext/nokogiri/native.h

@@ -9,6 +9,7 @@
 #include <libxml/xpathInternals.h>
 #include <libxml/xmlreader.h>
 #include <libxml/xmlsave.h>
+#include <libxml/xmlschemas.h>
 #include <libxml/HTMLparser.h>
 #include <libxml/HTMLtree.h>
 
@@ -88,6 +89,8 @@ int vasprintf (char **strp, const char *fmt, va_list ap);
 #include <html_sax_parser.h>
 #include <xslt_stylesheet.h>
 #include <xml_syntax_error.h>
+#include <xml_schema.h>
+#include <xml_relax_ng.h>
 
 extern VALUE mNokogiri ;
 extern VALUE mNokogiriXml ;

ext/nokogiri/xml_relax_ng.c

@@ -0,0 +1,102 @@
+#include <xml_relax_ng.h>
+
+static void dealloc(xmlRelaxNGPtr schema)
+{
+  NOKOGIRI_DEBUG_START(doc);
+  xmlRelaxNGFree(schema);
+  NOKOGIRI_DEBUG_END(doc);
+}
+
+/*
+ * call-seq:
+ *  validate_document(document)
+ *
+ * Validate a Nokogiri::XML::Document against this RelaxNG schema.
+ */
+static VALUE validate_document(VALUE self, VALUE document)
+{
+  xmlDocPtr doc;
+  xmlRelaxNGPtr schema;
+
+  Data_Get_Struct(self, xmlRelaxNG, schema);
+  Data_Get_Struct(document, xmlDoc, doc);
+
+  VALUE errors = rb_ary_new();
+
+  xmlRelaxNGValidCtxtPtr valid_ctxt = xmlRelaxNGNewValidCtxt(schema);
+
+  if(NULL == valid_ctxt) {
+    // we have a problem
+    rb_raise(rb_eRuntimeError, "Could not create a validation context");
+  }
+
+  xmlRelaxNGSetValidStructuredErrors(
+    valid_ctxt,
+    Nokogiri_error_array_pusher,
+    (void *)errors
+  );
+
+  xmlRelaxNGValidateDoc(valid_ctxt, doc);
+
+  xmlRelaxNGFreeValidCtxt(valid_ctxt);
+
+  return errors;
+}
+
+/*
+ * call-seq:
+ *  read_memory(string)
+ *
+ * Create a new RelaxNG from the contents of +string+
+ */
+static VALUE read_memory(VALUE klass, VALUE content)
+{
+  xmlRelaxNGParserCtxtPtr ctx = xmlRelaxNGNewMemParserCtxt(
+      (const char *)StringValuePtr(content),
+      RSTRING_LEN(content)
+  );
+
+  VALUE errors = rb_ary_new();
+  xmlSetStructuredErrorFunc((void *)errors, Nokogiri_error_array_pusher);
+
+  xmlRelaxNGSetParserStructuredErrors(
+      ctx,
+      Nokogiri_error_array_pusher,
+      (void *)errors
+  );
+
+  xmlRelaxNGPtr schema = xmlRelaxNGParse(ctx);
+
+  xmlSetStructuredErrorFunc(NULL, NULL);
+  xmlRelaxNGFreeParserCtxt(ctx);
+
+  if(NULL == schema) {
+    xmlErrorPtr error = xmlGetLastError();
+    if(error)
+      rb_funcall(rb_mKernel, rb_intern("raise"), 1,
+          Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error)
+      );
+    else
+      rb_raise(rb_eRuntimeError, "Could not parse document");
+
+    return Qnil;
+  }
+
+  VALUE rb_schema = Data_Wrap_Struct(klass, 0, dealloc, schema);
+  rb_iv_set(rb_schema, "@errors", errors);
+
+  return rb_schema;
+}
+
+VALUE cNokogiriXmlRelaxNG;
+void 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);
+
+  cNokogiriXmlRelaxNG = klass;
+
+  rb_define_singleton_method(klass, "read_memory", read_memory, 1);
+  rb_define_private_method(klass, "validate_document", validate_document, 1);
+}

ext/nokogiri/xml_relax_ng.h

@@ -0,0 +1,9 @@
+#ifndef NOKOGIRI_XML_RELAX_NG
+#define NOKOGIRI_XML_RELAX_NG
+
+#include <native.h>
+
+void init_xml_relax_ng();
+
+extern VALUE cNokogiriXmlRelaxNG;
+#endif

ext/nokogiri/xml_schema.c

@@ -0,0 +1,103 @@
+#include <xml_schema.h>
+
+static void dealloc(xmlSchemaPtr schema)
+{
+  NOKOGIRI_DEBUG_START(doc);
+  xmlSchemaFree(schema);
+  NOKOGIRI_DEBUG_END(doc);
+}
+
+/*
+ * call-seq:
+ *  validate_document(document)
+ *
+ * Validate a Nokogiri::XML::Document against this Schema.
+ */
+static VALUE validate_document(VALUE self, VALUE document)
+{
+  xmlDocPtr doc;
+  xmlSchemaPtr schema;
+
+  Data_Get_Struct(self, xmlSchema, schema);
+  Data_Get_Struct(document, xmlDoc, doc);
+
+  VALUE errors = rb_ary_new();
+
+  xmlSchemaValidCtxtPtr valid_ctxt = xmlSchemaNewValidCtxt(schema);
+
+  if(NULL == valid_ctxt) {
+    // we have a problem
+    rb_raise(rb_eRuntimeError, "Could not create a validation context");
+  }
+
+  xmlSchemaSetValidStructuredErrors(
+    valid_ctxt,
+    Nokogiri_error_array_pusher,
+    (void *)errors
+  );
+
+  xmlSchemaValidateDoc(valid_ctxt, doc);
+
+  xmlSchemaFreeValidCtxt(valid_ctxt);
+
+  return errors;
+}
+
+/*
+ * call-seq:
+ *  read_memory(string)
+ *
+ * Create a new Schema from the contents of +string+
+ */
+static VALUE read_memory(VALUE klass, VALUE content)
+{
+
+  xmlSchemaParserCtxtPtr ctx = xmlSchemaNewMemParserCtxt(
+      (const char *)StringValuePtr(content),
+      RSTRING_LEN(content)
+  );
+
+  VALUE errors = rb_ary_new();
+  xmlSetStructuredErrorFunc((void *)errors, Nokogiri_error_array_pusher);
+
+  xmlSchemaSetParserStructuredErrors(
+      ctx,
+      Nokogiri_error_array_pusher,
+      (void *)errors
+  );
+
+  xmlSchemaPtr schema = xmlSchemaParse(ctx);
+
+  xmlSetStructuredErrorFunc(NULL, NULL);
+  xmlSchemaFreeParserCtxt(ctx);
+
+  if(NULL == schema) {
+    xmlErrorPtr error = xmlGetLastError();
+    if(error)
+      rb_funcall(rb_mKernel, rb_intern("raise"), 1,
+          Nokogiri_wrap_xml_syntax_error((VALUE)NULL, error)
+      );
+    else
+      rb_raise(rb_eRuntimeError, "Could not parse document");
+
+    return Qnil;
+  }
+
+  VALUE rb_schema = Data_Wrap_Struct(klass, 0, dealloc, schema);
+  rb_iv_set(rb_schema, "@errors", errors);
+
+  return rb_schema;
+}
+
+VALUE cNokogiriXmlSchema;
+void init_xml_schema()
+{
+  VALUE nokogiri = rb_define_module("Nokogiri");
+  VALUE xml = rb_define_module_under(nokogiri, "XML");
+  VALUE klass = rb_define_class_under(xml, "Schema", rb_cObject);
+
+  cNokogiriXmlSchema = klass;
+
+  rb_define_singleton_method(klass, "read_memory", read_memory, 1);
+  rb_define_private_method(klass, "validate_document", validate_document, 1);
+}

ext/nokogiri/xml_schema.h

@@ -0,0 +1,9 @@
+#ifndef NOKOGIRI_XML_SCHEMA
+#define NOKOGIRI_XML_SCHEMA
+
+#include <native.h>
+
+void init_xml_schema();
+
+extern VALUE cNokogiriXmlSchema;
+#endif

lib/nokogiri/xml.rb

@@ -18,6 +18,8 @@
 require 'nokogiri/xml/notation'
 require 'nokogiri/xml/element'
 require 'nokogiri/xml/entity_declaration'
+require 'nokogiri/xml/schema'
+require 'nokogiri/xml/relax_ng'
 
 module Nokogiri
   class << self

lib/nokogiri/xml/relax_ng.rb

@@ -0,0 +1,29 @@
+module Nokogiri
+  module XML
+    class << self
+      def RelaxNG string_or_io
+        RelaxNG.new(string_or_io)
+      end
+    end
+
+    ###
+    # Nokogiri::XML::RelaxNG is used for validating XML against a
+    # RelaxNG schema.
+    #
+    # == Synopsis
+    #
+    # Validate an XML document against a RelaxNG schema.  Loop over the errors
+    # that are returned and print them out:
+    #
+    #   schema  = Nokogiri::XML::RelaxNG(File.read(ADDRESS_SCHEMA_FILE))
+    #   doc     = Nokogiri::XML(File.read(ADDRESS_XML_FILE))
+    #
+    #   schema.validate(doc).each do |error|
+    #     puts error.message
+    #   end
+    #
+    # The list of errors are Nokogiri::XML::SyntaxError objects.
+    class RelaxNG < Nokogiri::XML::Schema
+    end
+  end
+end