nokogiri 1.11.0.rc3-x86-linux → 1.11.0.rc4-x86-linux

data/ext/nokogiri/xml_xpath_context.c

@@ -1,6 +1,11 @@
 #include <xml_xpath_context.h>
 
-int vasprintf (char **strp, const char *fmt, va_list ap);
+/*
+ * these constants have matching declarations in
+ * ext/java/nokogiri/internals/NokogiriNamespaceContext.java
+ */
+static const xmlChar *NOKOGIRI_BUILTIN_PREFIX = (const xmlChar *)"nokogiri-builtin";
+static const xmlChar *NOKOGIRI_BUILTIN_URI = (const xmlChar *)"https://www.nokogiri.org/default_ns/ruby/builtins";
 
 static void deallocate(xmlXPathContextPtr ctx)
 {
@@ -9,6 +14,73 @@ static void deallocate(xmlXPathContextPtr ctx)
   NOKOGIRI_DEBUG_END(ctx);
 }
 
+/* find a CSS class in an HTML element's `class` attribute */
+const xmlChar* builtin_css_class(const xmlChar* str, const xmlChar *val)
+{
+  int val_len;
+
+  if (str == NULL) { return(NULL); }
+  if (val == NULL) { return(NULL); }
+
+  val_len = xmlStrlen(val);
+  if (val_len == 0) { return(str); }
+
+  while (*str != 0) {
+    if ((*str == *val) && !xmlStrncmp(str, val, val_len)) {
+      const xmlChar* next_byte = str + val_len;
+
+      /* only match if the next byte is whitespace or end of string */
+      if ((*next_byte == 0) || (IS_BLANK_CH(*next_byte))) {
+        return((const xmlChar*)str);
+      }
+    }
+
+    /* advance str to whitespace */
+    while ((*str != 0) && !IS_BLANK_CH(*str)) {
+      str++;
+    }
+
+    /* advance str to start of next word or end of string */
+    while ((*str != 0) && IS_BLANK_CH(*str)) {
+      str++;
+    }
+  }
+
+  return(NULL);
+}
+
+/* xmlXPathFunction to wrap builtin_css_class() */
+static void xpath_builtin_css_class(xmlXPathParserContextPtr ctxt, int nargs)
+{
+  xmlXPathObjectPtr hay, needle;
+
+  CHECK_ARITY(2);
+
+  CAST_TO_STRING;
+  needle = valuePop(ctxt);
+  if ((needle == NULL) || (needle->type != XPATH_STRING)) {
+    xmlXPathFreeObject(needle);
+    XP_ERROR(XPATH_INVALID_TYPE);
+  }
+
+  CAST_TO_STRING;
+  hay = valuePop(ctxt);
+  if ((hay == NULL) || (hay->type != XPATH_STRING)) {
+    xmlXPathFreeObject(hay);
+    xmlXPathFreeObject(needle);
+    XP_ERROR(XPATH_INVALID_TYPE);
+  }
+
+  if (builtin_css_class(hay->stringval, needle->stringval)) {
+    valuePush(ctxt, xmlXPathNewBoolean(1));
+  } else {
+    valuePush(ctxt, xmlXPathNewBoolean(0));
+  }
+
+  xmlXPathFreeObject(hay);
+  xmlXPathFreeObject(needle);
+}
+
 /*
  * call-seq:
  *  register_ns(prefix, uri)
@@ -263,14 +335,18 @@ static VALUE new(VALUE klass, VALUE nodeobj)
   xmlXPathContextPtr ctx;
   VALUE self;
 
-  xmlXPathInit();
-
   Data_Get_Struct(nodeobj, xmlNode, node);
 
+  xmlXPathInit();
+
   ctx = xmlXPathNewContext(node->doc);
   ctx->node = node;
+
+  xmlXPathRegisterNs(ctx, NOKOGIRI_BUILTIN_PREFIX, NOKOGIRI_BUILTIN_URI);
+  xmlXPathRegisterFuncNS(ctx, (const xmlChar *)"css-class", NOKOGIRI_BUILTIN_URI,
+                         xpath_builtin_css_class);
+
   self = Data_Wrap_Struct(klass, 0, deallocate, ctx);
-  /*rb_iv_set(self, "@xpath_handler", Qnil); */
   return self;
 }
 

data/ext/nokogiri/xslt_stylesheet.c

@@ -7,9 +7,6 @@
 
 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);
@@ -37,7 +34,7 @@ static void xslt_generic_error_handler(void * ctx, const char *msg, ...)
 
   rb_str_cat2((VALUE)ctx, message);
 
-  vasprintf_free(message);
+  free(message);
 }
 
 VALUE Nokogiri_wrap_xslt_stylesheet(xsltStylesheetPtr ss)

data/lib/nokogiri.rb

@@ -113,7 +113,7 @@ module Nokogiri
 end
 
 ###
-# Parser a document contained in +args+.  Nokogiri will try to guess what
+# Parse 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
 #

data/lib/nokogiri/css/parser.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 #
 # DO NOT MODIFY!!!!
-# This file is automatically generated by Racc 1.4.16
+# This file is automatically generated by Racc 1.5.1
 # from Racc grammar file "".
 #
 
@@ -476,7 +476,7 @@ def _reduce_26(val, _values, result)
 end
 
 def _reduce_27(val, _values, result)
-        # Non standard, but hpricot supports it.
+        # non-standard, from hpricot
         result = Node.new(:PSEUDO_CLASS,
           [Node.new(:FUNCTION, ['nth-child(', val[1]])]
         )
@@ -558,7 +558,7 @@ def _reduce_40(val, _values, result)
         when 'n'
           result = Node.new(:NTH, ['1','n','+','0'])
         else
-          # This is not CSS standard.  It allows us to support this:
+          # non-standard to support custom functions:
           # assert_xpath("//a[foo(., @href)]", @parser.parse('a:foo(@href)'))
           # assert_xpath("//a[foo(., @a, b)]", @parser.parse('a:foo(@a, b)'))
           # assert_xpath("//a[foo(., a, 10)]", @parser.parse('a:foo(a, 10)'))

data/lib/nokogiri/css/parser.y

@@ -88,7 +88,7 @@ rule
         )
       }
     | LSQUARE NUMBER RSQUARE {
-        # Non standard, but hpricot supports it.
+        # non-standard, from hpricot
         result = Node.new(:PSEUDO_CLASS,
           [Node.new(:FUNCTION, ['nth-child(', val[1]])]
         )
@@ -139,7 +139,7 @@ rule
         when 'n'
           result = Node.new(:NTH, ['1','n','+','0'])
         else
-          # This is not CSS standard.  It allows us to support this:
+          # non-standard to support custom functions:
           # assert_xpath("//a[foo(., @href)]", @parser.parse('a:foo(@href)'))
           # assert_xpath("//a[foo(., @a, b)]", @parser.parse('a:foo(@a, b)'))
           # assert_xpath("//a[foo(., a, 10)]", @parser.parse('a:foo(a, 10)'))

data/lib/nokogiri/css/xpath_visitor.rb

@@ -3,7 +3,6 @@ module Nokogiri
   module CSS
     class XPathVisitor # :nodoc:
       def visit_function node
-
         msg = :"visit_function_#{node.value.first.gsub(/[(]/, '')}"
         return self.send(msg, node) if self.respond_to?(msg)
 
@@ -13,50 +12,51 @@ module Nokogiri
         when /^self\(/
           "self::#{node.value[1]}"
         when /^eq\(/
-          "position() = #{node.value[1]}"
+          "position()=#{node.value[1]}"
         when /^(nth|nth-of-type)\(/
           if node.value[1].is_a?(Nokogiri::CSS::Node) and node.value[1].type == :NTH
             nth(node.value[1])
           else
-            "position() = #{node.value[1]}"
+            "position()=#{node.value[1]}"
           end
         when /^nth-child\(/
           if node.value[1].is_a?(Nokogiri::CSS::Node) and node.value[1].type == :NTH
             nth(node.value[1], :child => true)
           else
-            "count(preceding-sibling::*) = #{node.value[1].to_i-1}"
+            "count(preceding-sibling::*)=#{node.value[1].to_i-1}"
           end
         when /^nth-last-of-type\(/
           if node.value[1].is_a?(Nokogiri::CSS::Node) and node.value[1].type == :NTH
             nth(node.value[1], :last => true)
           else
             index = node.value[1].to_i - 1
-            index == 0 ? "position() = last()" : "position() = last() - #{index}"
+            index == 0 ? "position()=last()" : "position()=last()-#{index}"
           end
         when /^nth-last-child\(/
           if node.value[1].is_a?(Nokogiri::CSS::Node) and node.value[1].type == :NTH
             nth(node.value[1], :last => true, :child => true)
           else
-            "count(following-sibling::*) = #{node.value[1].to_i-1}"
+            "count(following-sibling::*)=#{node.value[1].to_i-1}"
           end
         when /^(first|first-of-type)\(/
-          "position() = 1"
+          "position()=1"
         when /^(last|last-of-type)\(/
-          "position() = last()"
+          "position()=last()"
         when /^contains\(/
-          "contains(., #{node.value[1]})"
+          "contains(.,#{node.value[1]})"
         when /^gt\(/
-          "position() > #{node.value[1]}"
+          "position()>#{node.value[1]}"
         when /^only-child\(/
-          "last() = 1"
+          "last()=1"
         when /^comment\(/
           "comment()"
         when /^has\(/
           is_direct = node.value[1].value[0].nil? # e.g. "has(> a)", "has(~ a)", "has(+ a)"
           ".#{"//" if !is_direct}#{node.value[1].accept(self)}"
         else
+          # non-standard. this looks like a function call.
           args = ['.'] + node.value[1..-1]
-          "#{node.value.first}#{args.join(', ')})"
+          "#{node.value.first}#{args.join(',')})"
         end
       end
 
@@ -71,18 +71,18 @@ module Nokogiri
 
       def visit_id node
         node.value.first =~ /^#(.*)$/
-        "@id = '#{$1}'"
+        "@id='#{$1}'"
       end
 
       def visit_attribute_condition node
-         attribute = if (node.value.first.type == :FUNCTION) or (node.value.first.value.first =~ /::/)
-                       ''
-                     else
-                       '@'
-                     end
+        attribute = if (node.value.first.type == :FUNCTION) or (node.value.first.value.first =~ /::/)
+                      ''
+                    else
+                      '@'
+                    end
         attribute += node.value.first.accept(self)
 
-        # Support non-standard css
+        # non-standard. attributes starting with '@'
         attribute.gsub!(/^@@/, '@')
 
         return attribute unless node.value.length == 3
@@ -90,29 +90,30 @@ module Nokogiri
         value = node.value.last
         value = "'#{value}'" if value !~ /^['"]/
 
+        # quoted values - see test_attribute_value_with_quotes in test/css/test_parser.rb
         if (value[0]==value[-1]) && %q{"'}.include?(value[0])
           str_value = value[1..-2]
           if str_value.include?(value[0])
-            value = 'concat("' + str_value.split('"', -1).join(%q{", '"', "}) + '", "")'
+            value = 'concat("' + str_value.split('"', -1).join(%q{",'"',"}) + '","")'
           end
         end
 
         case node.value[1]
         when :equal
-          attribute + " = " + "#{value}"
+          attribute + "=" + "#{value}"
         when :not_equal
-          attribute + " != " + "#{value}"
+          attribute + "!=" + "#{value}"
         when :substring_match
-          "contains(#{attribute}, #{value})"
+          "contains(#{attribute},#{value})"
         when :prefix_match
-          "starts-with(#{attribute}, #{value})"
+          "starts-with(#{attribute},#{value})"
         when :dash_match
-          "#{attribute} = #{value} or starts-with(#{attribute}, concat(#{value}, '-'))"
+          "#{attribute}=#{value} or starts-with(#{attribute},concat(#{value},'-'))"
         when :includes
-          "contains(concat(\" \", #{attribute}, \" \"),concat(\" \", #{value}, \" \"))"
+          value = value[1..-2] # strip quotes
+          css_class(attribute, value)
         when :suffix_match
-          "substring(#{attribute}, string-length(#{attribute}) - " +
-            "string-length(#{value}) + 1, string-length(#{value})) = #{value}"
+          "substring(#{attribute},string-length(#{attribute})-string-length(#{value})+1,string-length(#{value}))=#{value}"
         else
           attribute + " #{node.value[1]} " + "#{value}"
         end
@@ -126,14 +127,14 @@ module Nokogiri
           return self.send(msg, node) if self.respond_to?(msg)
 
           case node.value.first
-          when "first" then "position() = 1"
-          when "first-child" then "count(preceding-sibling::*) = 0"
-          when "last" then "position() = last()"
-          when "last-child" then "count(following-sibling::*) = 0"
-          when "first-of-type" then "position() = 1"
-          when "last-of-type" then "position() = last()"
-          when "only-child" then "count(preceding-sibling::*) = 0 and count(following-sibling::*) = 0"
-          when "only-of-type" then "last() = 1"
+          when "first" then "position()=1"
+          when "first-child" then "count(preceding-sibling::*)=0"
+          when "last" then "position()=last()"
+          when "last-child" then "count(following-sibling::*)=0"
+          when "first-of-type" then "position()=1"
+          when "last-of-type" then "position()=last()"
+          when "only-child" then "count(preceding-sibling::*)=0 and count(following-sibling::*)=0"
+          when "only-of-type" then "last()=1"
           when "empty" then "not(node())"
           when "parent" then "node()"
           when "root" then "not(parent::*)"
@@ -144,7 +145,7 @@ module Nokogiri
       end
 
       def visit_class_condition node
-        "contains(concat(' ', normalize-space(@class), ' '), ' #{node.value.first} ')"
+        css_class("@class", node.value.first)
       end
 
       def visit_combinator node
@@ -181,25 +182,26 @@ module Nokogiri
         node.accept(self)
       end
 
-    private
+      private
+
       def nth node, options={}
         raise ArgumentError, "expected an+b node to contain 4 tokens, but is #{node.value.inspect}" unless node.value.size == 4
 
         a, b = read_a_and_positive_b node.value
         position = if options[:child]
-          options[:last] ? "(count(following-sibling::*) + 1)" : "(count(preceding-sibling::*) + 1)"
+          options[:last] ? "(count(following-sibling::*)+1)" : "(count(preceding-sibling::*)+1)"
         else
           options[:last] ? "(last()-position()+1)" : "position()"
         end
 
         if b.zero?
-          "(#{position} mod #{a}) = 0"
+          "(#{position} mod #{a})=0"
         else
           compare = a < 0 ? "<=" : ">="
           if a.abs == 1
-            "#{position} #{compare} #{b}"
+            "#{position}#{compare}#{b}"
           else
-            "(#{position} #{compare} #{b}) and (((#{position}-#{b}) mod #{a.abs}) = 0)"
+            "(#{position}#{compare}#{b}) and (((#{position}-#{b}) mod #{a.abs})=0)"
           end
         end
       end
@@ -227,6 +229,32 @@ module Nokogiri
           end =~ /(nth|first|last|only)-of-type(\()?/
         end
       end
+
+      # use only ordinary xpath functions
+      def css_class_standard(hay, needle)
+        "contains(concat(' ',normalize-space(#{hay}),' '),' #{needle} ')"
+      end
+
+      # use the builtin implementation
+      def css_class_builtin(hay, needle)
+        "nokogiri-builtin:css-class(#{hay},'#{needle}')"
+      end
+
+      alias_method :css_class, :css_class_standard
+    end
+
+    class XPathVisitorAlwaysUseBuiltins < XPathVisitor # :nodoc:
+      private
+      alias_method :css_class, :css_class_builtin
+    end
+
+    class XPathVisitorOptimallyUseBuiltins < XPathVisitor # :nodoc:
+      private
+      if Nokogiri.uses_libxml?
+        alias_method :css_class, :css_class_builtin
+      else
+        alias_method :css_class, :css_class_standard
+      end
     end
   end
 end

data/lib/nokogiri/html/document.rb

@@ -1,4 +1,7 @@
 # frozen_string_literal: true
+
+require 'pathname'
+
 module Nokogiri
   module HTML
     class Document < Nokogiri::XML::Document
@@ -161,11 +164,12 @@ module Nokogiri
         # Nokogiri::XML::ParseOptions::RECOVER.  See the constants in
         # Nokogiri::XML::ParseOptions.
         def parse string_or_io, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_HTML
-
           options = Nokogiri::XML::ParseOptions.new(options) if Integer === options
-          # Give the options to the user
+
           yield options if block_given?
 
+          url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
+
           if string_or_io.respond_to?(:encoding)
             unless string_or_io.encoding.name == "ASCII-8BIT"
               encoding ||= string_or_io.encoding.name
@@ -173,7 +177,12 @@ module Nokogiri
           end
 
           if string_or_io.respond_to?(:read)
-            url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
+            if string_or_io.is_a?(Pathname)
+              # resolve the Pathname to the file and open it as an IO object, see #2110
+              string_or_io = string_or_io.expand_path.open
+              url ||= string_or_io.path
+            end
+
             unless encoding
               # Libxml2's parser has poor support for encoding
               # detection.  First, it does not recognize the HTML5
@@ -252,9 +261,6 @@ module Nokogiri
         end
 
         def self.detect_encoding(chunk)
-          if Nokogiri.jruby? && EncodingReader.is_jruby_without_fix?
-            return EncodingReader.detect_encoding_for_jruby_without_fix(chunk)
-          end
           m = chunk.match(/\A(<\?xml[ \t\r\n]+[^>]*>)/) and
             return Nokogiri.XML(m[1]).encoding
 
@@ -273,26 +279,6 @@ module Nokogiri
           end
         end
 
-        def self.is_jruby_without_fix?
-          JRUBY_VERSION.split('.').join.to_i < 165
-        end
-
-        def self.detect_encoding_for_jruby_without_fix(chunk)
-          m = chunk.match(/\A(<\?xml[ \t\r\n]+[^>]*>)/) and
-            return Nokogiri.XML(m[1]).encoding
-
-          m = chunk.match(/(<meta\s)(.*)(charset\s*=\s*([\w-]+))(.*)/i) and
-            return m[4]
-
-          catch(:encoding_found) {
-            Nokogiri::HTML::SAX::Parser.new(JumpSAXHandler.new(:encoding_found.to_s)).parse(chunk)
-            nil
-          }
-        rescue Nokogiri::SyntaxError, RuntimeError
-          # Ignore parser errors that nokogiri may raise
-          nil
-        end
-
         def initialize(io)
           @io = io
           @firstchunk = nil