nokogiri 1.11.0.rc1-x86-linux

data/ext/nokogiri/xml_xpath_context.h

@@ -0,0 +1,10 @@
+#ifndef NOKOGIRI_XML_XPATH_CONTEXT
+#define NOKOGIRI_XML_XPATH_CONTEXT
+
+#include <nokogiri.h>
+
+void init_xml_xpath_context();
+void Nokogiri_marshal_xpath_funcall_and_return_values(xmlXPathParserContextPtr ctx, int nargs, VALUE handler, const char* function_name) ;
+
+extern VALUE cNokogiriXmlXpathContext;
+#endif

data/ext/nokogiri/xslt_stylesheet.c

@@ -0,0 +1,266 @@
+#include <xslt_stylesheet.h>
+
+#include <libxslt/xsltInternals.h>
+#include <libxslt/xsltutils.h>
+#include <libxslt/transform.h>
+#include <libexslt/exslt.h>
+
+VALUE xslt;
+
+int vasprintf (char **strp, const char *fmt, va_list ap);
+void vasprintf_free (void *p);
+
+static void mark(nokogiriXsltStylesheetTuple *wrapper)
+{
+  rb_gc_mark(wrapper->func_instances);
+}
+
+static void dealloc(nokogiriXsltStylesheetTuple *wrapper)
+{
+    xsltStylesheetPtr doc = wrapper->ss;
+
+    NOKOGIRI_DEBUG_START(doc);
+    xsltFreeStylesheet(doc); /* commented out for now. */
+    NOKOGIRI_DEBUG_END(doc);
+
+    free(wrapper);
+}
+
+static void xslt_generic_error_handler(void * ctx, const char *msg, ...)
+{
+  char * message;
+
+  va_list args;
+  va_start(args, msg);
+  vasprintf(&message, msg, args);
+  va_end(args);
+
+  rb_str_cat2((VALUE)ctx, message);
+
+  vasprintf_free(message);
+}
+
+VALUE Nokogiri_wrap_xslt_stylesheet(xsltStylesheetPtr ss)
+{
+  VALUE self;
+  nokogiriXsltStylesheetTuple *wrapper;
+
+  self = Data_Make_Struct(cNokogiriXsltStylesheet, nokogiriXsltStylesheetTuple,
+                          mark, dealloc, wrapper);
+
+  ss->_private = (void *)self;
+  wrapper->ss = ss;
+  wrapper->func_instances = rb_ary_new();
+
+  return self;
+}
+
+/*
+ * call-seq:
+ *   parse_stylesheet_doc(document)
+ *
+ * Parse a stylesheet from +document+.
+ */
+static VALUE parse_stylesheet_doc(VALUE klass, VALUE xmldocobj)
+{
+    xmlDocPtr xml, xml_cpy;
+    VALUE errstr, exception;
+    xsltStylesheetPtr ss ;
+    Data_Get_Struct(xmldocobj, xmlDoc, xml);
+    exsltRegisterAll();
+
+    errstr = rb_str_new(0, 0);
+    xsltSetGenericErrorFunc((void *)errstr, xslt_generic_error_handler);
+
+    xml_cpy = xmlCopyDoc(xml, 1); /* 1 => recursive */
+    ss = xsltParseStylesheetDoc(xml_cpy);
+
+    xsltSetGenericErrorFunc(NULL, NULL);
+
+    if (!ss) {
+	xmlFreeDoc(xml_cpy);
+	exception = rb_exc_new3(rb_eRuntimeError, errstr);
+	rb_exc_raise(exception);
+    }
+
+    return Nokogiri_wrap_xslt_stylesheet(ss);
+}
+
+
+/*
+ * call-seq:
+ *   serialize(document)
+ *
+ * Serialize +document+ to an xml string.
+ */
+static VALUE serialize(VALUE self, VALUE xmlobj)
+{
+    xmlDocPtr xml ;
+    nokogiriXsltStylesheetTuple *wrapper;
+    xmlChar* doc_ptr ;
+    int doc_len ;
+    VALUE rval ;
+
+    Data_Get_Struct(xmlobj, xmlDoc, xml);
+    Data_Get_Struct(self, nokogiriXsltStylesheetTuple, wrapper);
+    xsltSaveResultToString(&doc_ptr, &doc_len, xml, wrapper->ss);
+    rval = NOKOGIRI_STR_NEW(doc_ptr, doc_len);
+    xmlFree(doc_ptr);
+    return rval ;
+}
+
+/*
+ *  call-seq:
+ *    transform(document, params = [])
+ *
+ *  Apply an XSLT stylesheet to an XML::Document.
+ *  +params+ is an array of strings used as XSLT parameters.
+ *  returns Nokogiri::XML::Document
+ *
+ *  Example:
+ *
+ *    doc   = Nokogiri::XML(File.read(ARGV[0]))
+ *    xslt  = Nokogiri::XSLT(File.read(ARGV[1]))
+ *    puts xslt.transform(doc, ['key', 'value'])
+ *
+ */
+static VALUE transform(int argc, VALUE* argv, VALUE self)
+{
+    VALUE xmldoc, paramobj, errstr, exception ;
+    xmlDocPtr xml ;
+    xmlDocPtr result ;
+    nokogiriXsltStylesheetTuple *wrapper;
+    const char** params ;
+    long param_len, j ;
+    int parse_error_occurred ;
+
+    rb_scan_args(argc, argv, "11", &xmldoc, &paramobj);
+    if (NIL_P(paramobj)) { paramobj = rb_ary_new2(0L) ; }
+    if (!rb_obj_is_kind_of(xmldoc, cNokogiriXmlDocument))
+      rb_raise(rb_eArgError, "argument must be a Nokogiri::XML::Document");
+
+    /* handle hashes as arguments. */
+    if(T_HASH == TYPE(paramobj)) {
+      paramobj = rb_funcall(paramobj, rb_intern("to_a"), 0);
+      paramobj = rb_funcall(paramobj, rb_intern("flatten"), 0);
+    }
+
+    Check_Type(paramobj, T_ARRAY);
+
+    Data_Get_Struct(xmldoc, xmlDoc, xml);
+    Data_Get_Struct(self, nokogiriXsltStylesheetTuple, wrapper);
+
+    param_len = RARRAY_LEN(paramobj);
+    params = calloc((size_t)param_len+1, sizeof(char*));
+    for (j = 0 ; j < param_len ; j++) {
+      VALUE entry = rb_ary_entry(paramobj, j);
+      const char * ptr = StringValueCStr(entry);
+      params[j] = ptr;
+    }
+    params[param_len] = 0 ;
+
+    errstr = rb_str_new(0, 0);
+    xsltSetGenericErrorFunc((void *)errstr, xslt_generic_error_handler);
+    xmlSetGenericErrorFunc((void *)errstr, xslt_generic_error_handler);
+
+    result = xsltApplyStylesheet(wrapper->ss, xml, params);
+    free(params);
+
+    xsltSetGenericErrorFunc(NULL, NULL);
+    xmlSetGenericErrorFunc(NULL, NULL);
+
+    parse_error_occurred = (Qfalse == rb_funcall(errstr, rb_intern("empty?"), 0));
+
+    if (parse_error_occurred) {
+      exception = rb_exc_new3(rb_eRuntimeError, errstr);
+      rb_exc_raise(exception);
+    }
+
+    return Nokogiri_wrap_xml_document((VALUE)0, result) ;
+}
+
+static void method_caller(xmlXPathParserContextPtr ctxt, int nargs)
+{
+    VALUE handler;
+    const char *function_name;
+    xsltTransformContextPtr transform;
+    const xmlChar *functionURI;
+
+    transform = xsltXPathGetTransformContext(ctxt);
+    functionURI = ctxt->context->functionURI;
+    handler = (VALUE)xsltGetExtData(transform, functionURI);
+    function_name = (const char*)(ctxt->context->function);
+
+    Nokogiri_marshal_xpath_funcall_and_return_values(ctxt, nargs, handler, (const char*)function_name);
+}
+
+static void * initFunc(xsltTransformContextPtr ctxt, const xmlChar *uri)
+{
+    VALUE modules = rb_iv_get(xslt, "@modules");
+    VALUE obj = rb_hash_aref(modules, rb_str_new2((const char *)uri));
+    VALUE args = { Qfalse };
+    VALUE methods = rb_funcall(obj, rb_intern("instance_methods"), 1, args);
+    VALUE inst;
+    nokogiriXsltStylesheetTuple *wrapper;
+    int i;
+
+    for(i = 0; i < RARRAY_LEN(methods); i++) {
+	VALUE method_name = rb_obj_as_string(rb_ary_entry(methods, i));
+	xsltRegisterExtFunction(ctxt,
+          (unsigned char *)StringValueCStr(method_name), uri, method_caller);
+    }
+
+    Data_Get_Struct((VALUE)ctxt->style->_private, nokogiriXsltStylesheetTuple,
+                    wrapper);
+    inst = rb_class_new_instance(0, NULL, obj);
+    rb_ary_push(wrapper->func_instances, inst);
+
+    return (void *)inst;
+}
+
+static void shutdownFunc(xsltTransformContextPtr ctxt,
+	const xmlChar *uri, void *data)
+{
+    nokogiriXsltStylesheetTuple *wrapper;
+
+    Data_Get_Struct((VALUE)ctxt->style->_private, nokogiriXsltStylesheetTuple,
+                    wrapper);
+
+    rb_ary_clear(wrapper->func_instances);
+}
+
+/*
+ *  call-seq:
+ *    register(uri, custom_handler_class)
+ *
+ *  Register a class that implements custom XSLT transformation functions.
+ */
+static VALUE registr(VALUE self, VALUE uri, VALUE obj)
+{
+    VALUE modules = rb_iv_get(self, "@modules");
+    if(NIL_P(modules)) rb_raise(rb_eRuntimeError, "wtf! @modules isn't set");
+
+    rb_hash_aset(modules, uri, obj);
+    xsltRegisterExtModule((unsigned char *)StringValueCStr(uri), initFunc, shutdownFunc);
+    return self;
+}
+
+VALUE cNokogiriXsltStylesheet ;
+void init_xslt_stylesheet()
+{
+  VALUE nokogiri;
+  VALUE klass;
+
+  nokogiri = rb_define_module("Nokogiri");
+  xslt = rb_define_module_under(nokogiri, "XSLT");
+  klass = rb_define_class_under(xslt, "Stylesheet", rb_cObject);
+
+  rb_iv_set(xslt, "@modules", rb_hash_new());
+
+  cNokogiriXsltStylesheet = klass;
+
+  rb_define_singleton_method(klass, "parse_stylesheet_doc", parse_stylesheet_doc, 1);
+  rb_define_singleton_method(xslt, "register", registr, 2);
+  rb_define_method(klass, "serialize", serialize, 1);
+  rb_define_method(klass, "transform", transform, -1);
+}

data/ext/nokogiri/xslt_stylesheet.h

@@ -0,0 +1,14 @@
+#ifndef NOKOGIRI_XSLT_STYLESHEET
+#define NOKOGIRI_XSLT_STYLESHEET
+
+#include <nokogiri.h>
+
+void init_xslt_stylesheet();
+
+extern VALUE cNokogiriXsltStylesheet ;
+
+typedef struct _nokogiriXsltStylesheetTuple {
+  xsltStylesheetPtr ss;
+  VALUE func_instances;
+} nokogiriXsltStylesheetTuple;
+#endif

data/lib/nokogiri.rb

@@ -0,0 +1,127 @@
+# -*- coding: utf-8 -*-
+# frozen_string_literal: true
+# Modify the PATH on windows so that the external DLLs will get loaded.
+
+require 'rbconfig'
+
+if defined?(RUBY_ENGINE) && RUBY_ENGINE == "jruby"
+  require 'nokogiri/jruby/dependencies'
+end
+
+begin
+  RUBY_VERSION =~ /(\d+\.\d+)/
+  require "nokogiri/#{$1}/nokogiri"
+rescue LoadError
+  require 'nokogiri/nokogiri'
+end
+require 'nokogiri/version'
+require 'nokogiri/syntax_error'
+require 'nokogiri/xml'
+require 'nokogiri/xslt'
+require 'nokogiri/html'
+require 'nokogiri/decorators/slop'
+require 'nokogiri/css'
+require 'nokogiri/html/builder'
+
+# Nokogiri parses and searches XML/HTML very quickly, and also has
+# correctly implemented CSS3 selector support as well as XPath 1.0
+# support.
+#
+# Parsing a document returns either a Nokogiri::XML::Document, or a
+# Nokogiri::HTML::Document depending on the kind of document you parse.
+#
+# Here is an example:
+#
+#   require 'nokogiri'
+#   require 'open-uri'
+#
+#   # Get a Nokogiri::HTML:Document for the page we’re interested in...
+#
+#   doc = Nokogiri::HTML(open('http://www.google.com/search?q=tenderlove'))
+#
+#   # Do funky things with it using Nokogiri::XML::Node methods...
+#
+#   ####
+#   # Search for nodes by css
+#   doc.css('h3.r a.l').each do |link|
+#     puts link.content
+#   end
+#
+# See Nokogiri::XML::Searchable#css for more information about CSS searching.
+# See Nokogiri::XML::Searchable#xpath for more information about XPath searching.
+module Nokogiri
+  class << self
+    ###
+    # Parse an HTML or XML document.  +string+ contains the document.
+    def parse string, url = nil, encoding = nil, options = nil
+      if string.respond_to?(:read) ||
+          /^\s*<(?:!DOCTYPE\s+)?html[\s>]/i === string[0, 512]
+        # Expect an HTML indicator to appear within the first 512
+        # characters of a document. (<?xml ?> + <?xml-stylesheet ?>
+        # shouldn't be that long)
+        Nokogiri.HTML(string, url, encoding,
+          options || XML::ParseOptions::DEFAULT_HTML)
+      else
+        Nokogiri.XML(string, url, encoding,
+          options || XML::ParseOptions::DEFAULT_XML)
+      end.tap { |doc|
+        yield doc if block_given?
+      }
+    end
+
+    ###
+    # Create a new Nokogiri::XML::DocumentFragment
+    def make input = nil, opts = {}, &blk
+      if input
+        Nokogiri::HTML.fragment(input).children.first
+      else
+        Nokogiri(&blk)
+      end
+    end
+
+    ###
+    # Parse a document and add the Slop decorator.  The Slop decorator
+    # implements method_missing such that methods may be used instead of CSS
+    # or XPath.  For example:
+    #
+    #   doc = Nokogiri::Slop(<<-eohtml)
+    #     <html>
+    #       <body>
+    #         <p>first</p>
+    #         <p>second</p>
+    #       </body>
+    #     </html>
+    #   eohtml
+    #   assert_equal('second', doc.html.body.p[1].text)
+    #
+    def Slop(*args, &block)
+      Nokogiri(*args, &block).slop!
+    end
+
+    def install_default_aliases
+      # Make sure to support some popular encoding aliases not known by
+      # all iconv implementations.
+      {
+        'Windows-31J' => 'CP932',	# Windows-31J is the IANA registered name of CP932.
+      }.each { |alias_name, name|
+        EncodingHandler.alias(name, alias_name) if EncodingHandler[alias_name].nil?
+      }
+    end
+  end
+
+  Nokogiri.install_default_aliases
+end
+
+###
+# Parser a document contained in +args+.  Nokogiri will try to guess what
+# type of document you are attempting to parse.  For more information, see
+# Nokogiri.parse
+#
+# To specify the type of document, use Nokogiri.XML or Nokogiri.HTML.
+def Nokogiri(*args, &block)
+  if block_given?
+    Nokogiri::HTML::Builder.new(&block).doc.root
+  else
+    Nokogiri.parse(*args)
+  end
+end

data/lib/nokogiri/css.rb

@@ -0,0 +1,28 @@
+# frozen_string_literal: true
+require 'nokogiri/css/node'
+require 'nokogiri/css/xpath_visitor'
+x = $-w
+$-w = false
+require 'nokogiri/css/parser'
+$-w = x
+
+require 'nokogiri/css/tokenizer'
+require 'nokogiri/css/syntax_error'
+
+module Nokogiri
+  module CSS
+    class << self
+      ###
+      # Parse this CSS selector in +selector+.  Returns an AST.
+      def parse selector
+        Parser.new.parse selector
+      end
+
+      ###
+      # Get the XPath for +selector+.
+      def xpath_for selector, options={}
+        Parser.new(options[:ns] || {}).xpath_for selector, options
+      end
+    end
+  end
+end