nokogiri 1.10.9 → 1.12.5

data/lib/nokogiri/css/xpath_visitor.rb

@@ -1,8 +1,8 @@
+# frozen_string_literal: true
 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)
 
@@ -12,49 +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\(/
-          ".//#{node.value[1].accept(self)}"
+          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
 
@@ -69,18 +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
@@ -88,29 +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
@@ -124,14 +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::*)"
@@ -142,7 +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
@@ -179,25 +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
@@ -225,6 +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/css.rb

@@ -1,27 +1,28 @@
-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'
-
+# frozen_string_literal: true
 module Nokogiri
   module CSS
     class << self
       ###
       # Parse this CSS selector in +selector+.  Returns an AST.
-      def parse selector
-        Parser.new.parse selector
+      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
+      def xpath_for(selector, options = {})
+        Parser.new(options[:ns] || {}).xpath_for(selector, options)
       end
     end
   end
 end
+
+require_relative "css/node"
+require_relative "css/xpath_visitor"
+x = $-w
+$-w = false
+require_relative "css/parser"
+$-w = x
+
+require_relative "css/tokenizer"
+require_relative "css/syntax_error"

data/lib/nokogiri/decorators/slop.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 module Nokogiri
   module Decorators
     ###

data/lib/nokogiri/extension.rb

@@ -0,0 +1,31 @@
+# frozen_string_literal: true
+
+# load the C or Java extension
+begin
+  # native precompiled gems package shared libraries in <gem_dir>/lib/nokogiri/<ruby_version>
+  ::RUBY_VERSION =~ /(\d+\.\d+)/
+  require_relative "#{Regexp.last_match(1)}/nokogiri"
+rescue LoadError => e
+  if e.message =~ /GLIBC/
+    warn(<<~EOM)
+
+      ERROR: It looks like you're trying to use Nokogiri as a precompiled native gem on a system with glibc < 2.17:
+
+        #{e.message}
+
+        If that's the case, then please install Nokogiri via the `ruby` platform gem:
+            gem install nokogiri --platform=ruby
+        or:
+            bundle config set force_ruby_platform true
+
+        Please visit https://nokogiri.org/tutorials/installing_nokogiri.html for more help.
+
+    EOM
+    raise e
+  end
+
+  # use "require" instead of "require_relative" because non-native gems will place C extension files
+  # in Gem::BasicSpecification#extension_dir after compilation (during normal installation), which
+  # is in $LOAD_PATH but not necessarily relative to this file (see #2300)
+  require "nokogiri/nokogiri"
+end

data/lib/nokogiri/gumbo.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+module Nokogiri
+  module Gumbo
+    # The default maximum number of attributes per element.
+    DEFAULT_MAX_ATTRIBUTES = 400
+
+    # The default maximum number of errors for parsing a document or a fragment.
+    DEFAULT_MAX_ERRORS = 0
+
+    # The default maximum depth of the DOM tree produced by parsing a document
+    # or fragment.
+    DEFAULT_MAX_TREE_DEPTH = 400
+  end
+end

data/lib/nokogiri/html.rb

@@ -1,37 +1,42 @@
-require 'nokogiri/html/entity_lookup'
-require 'nokogiri/html/document'
-require 'nokogiri/html/document_fragment'
-require 'nokogiri/html/sax/parser_context'
-require 'nokogiri/html/sax/parser'
-require 'nokogiri/html/sax/push_parser'
-require 'nokogiri/html/element_description'
-require 'nokogiri/html/element_description_defaults'
+# frozen_string_literal: true
+require_relative "html4"
 
 module Nokogiri
-  class << self
-    ###
-    # Parse HTML.  Convenience method for Nokogiri::HTML::Document.parse
-    def HTML thing, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_HTML, &block
-      Nokogiri::HTML::Document.parse(thing, url, encoding, options, &block)
-    end
-  end
+  HTML = Nokogiri::HTML4
+
+  # @!method HTML(input, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_HTML, &block)
+  # Parse HTML. Convenience method for Nokogiri::HTML4::Document.parse
+  # @!scope class
+  define_singleton_method(:HTML, Nokogiri.method(:HTML4))
 
+  # @note This module/namespace is an alias for {Nokogiri::HTML4} as of v1.12.0. Before v1.12.0,
+  #   {Nokogiri::HTML4} did not exist, and this was the module/namespace for all HTML-related
+  #   classes.
   module HTML
-    class << self
-      ###
-      # Parse HTML.  Convenience method for Nokogiri::HTML::Document.parse
-      def parse thing, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_HTML, &block
-        Document.parse(thing, url, encoding, options, &block)
+    # @note This class is an alias for {Nokogiri::HTML4::Document} as of v1.12.0.
+    class Document < Nokogiri::XML::Document
+    end
+
+    # @note This class is an alias for {Nokogiri::HTML4::DocumentFragment} as of v1.12.0.
+    class DocumentFragment < Nokogiri::XML::DocumentFragment
+    end
+
+    # @note This class is an alias for {Nokogiri::HTML4::Builder} as of v1.12.0.
+    class Builder < Nokogiri::XML::Builder
+    end
+
+    module SAX
+      # @note This class is an alias for {Nokogiri::HTML4::SAX::Parser} as of v1.12.0.
+      class Parser < Nokogiri::XML::SAX::Parser
       end
 
-      ####
-      # Parse a fragment from +string+ in to a NodeSet.
-      def fragment string, encoding = nil
-        HTML::DocumentFragment.parse string, encoding
+      # @note This class is an alias for {Nokogiri::HTML4::SAX::ParserContext} as of v1.12.0.
+      class ParserContext < Nokogiri::XML::SAX::ParserContext
       end
-    end
 
-    # Instance of Nokogiri::HTML::EntityLookup
-    NamedCharacters = EntityLookup.new
+      # @note This class is an alias for {Nokogiri::HTML4::SAX::PushParser} as of v1.12.0.
+      class PushParser
+      end
+    end
   end
 end

data/lib/nokogiri/{html → html4}/builder.rb

@@ -1,5 +1,6 @@
+# frozen_string_literal: true
 module Nokogiri
-  module HTML
+  module HTML4
     ###
     # Nokogiri HTML builder is used for building HTML documents.  It is very
     # similar to the Nokogiri::XML::Builder.  In fact, you should go read the
@@ -11,7 +12,7 @@ module Nokogiri
     # Create an HTML document with a body that has an onload attribute, and a
     # span tag with a class of "bold" that has content of "Hello world".
     #
-    #   builder = Nokogiri::HTML::Builder.new do |doc|
+    #   builder = Nokogiri::HTML4::Builder.new do |doc|
     #     doc.html {
     #       doc.body(:onload => 'some_func();') {
     #         doc.span.bold {

data/lib/nokogiri/{html → html4}/document.rb

@@ -1,5 +1,9 @@
+# frozen_string_literal: true
+
+require 'pathname'
+
 module Nokogiri
-  module HTML
+  module HTML4
     class Document < Nokogiri::XML::Document
       ###
       # Get the meta tag encoding for this document.  If there is no meta tag,
@@ -160,11 +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
@@ -172,7 +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
@@ -251,9 +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
 
@@ -261,37 +268,17 @@ module Nokogiri
             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)).parse(chunk)
+              Nokogiri::HTML4::SAX::Parser.new(JumpSAXHandler.new(:encoding_found)).parse(chunk)
               nil
             }
           else
             handler = SAXHandler.new
-            parser = Nokogiri::HTML::SAX::PushParser.new(handler)
+            parser = Nokogiri::HTML4::SAX::PushParser.new(handler)
             parser << chunk rescue Nokogiri::SyntaxError
             handler.encoding
           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
@@ -299,7 +286,7 @@ module Nokogiri
         end
 
         # This method is used by the C extension so that
-        # Nokogiri::HTML::Document#read_io() does not leak memory when
+        # Nokogiri::HTML4::Document#read_io() does not leak memory when
         # EncodingFound is raised.
         attr_reader :encoding_found
 

data/lib/nokogiri/{html → html4}/document_fragment.rb

@@ -1,28 +1,29 @@
+# frozen_string_literal: true
 module Nokogiri
-  module HTML
+  module HTML4
     class DocumentFragment < Nokogiri::XML::DocumentFragment
       ####
       # Create a Nokogiri::XML::DocumentFragment from +tags+, using +encoding+
-      def self.parse tags, encoding = nil
-        doc = HTML::Document.new
+      def self.parse(tags, encoding = nil)
+        doc = HTML4::Document.new
 
         encoding ||= if tags.respond_to?(:encoding)
-                       encoding = tags.encoding
-                       if encoding == ::Encoding::ASCII_8BIT
-                         'UTF-8'
-                       else
-                         encoding.name
-                       end
-                     else
-                       'UTF-8'
-                     end
+          encoding = tags.encoding
+          if encoding == ::Encoding::ASCII_8BIT
+            'UTF-8'
+          else
+            encoding.name
+          end
+        else
+          'UTF-8'
+        end
 
         doc.encoding = encoding
 
         new(doc, tags)
       end
 
-      def initialize document, tags = nil, ctx = nil
+      def initialize(document, tags = nil, ctx = nil)
         return self unless tags
 
         if ctx
@@ -32,13 +33,13 @@ module Nokogiri
           self.errors = document.errors - preexisting_errors
         else
           # This is a horrible hack, but I don't care
-          if tags.strip =~ /^<body/i
-            path = "/html/body"
+          path = if /^\s*?<body/i.match?(tags)
+            "/html/body"
           else
-            path = "/html/body/node()"
+            "/html/body/node()"
           end
 
-          temp_doc = HTML::Document.parse "<html><body>#{tags}", nil, document.encoding
+          temp_doc = HTML4::Document.parse("<html><body>#{tags}", nil, document.encoding)
           temp_doc.xpath(path).each { |child| child.parent = self }
           self.errors = temp_doc.errors
         end

data/lib/nokogiri/{html → html4}/element_description.rb

@@ -1,5 +1,6 @@
+# frozen_string_literal: true
 module Nokogiri
-  module HTML
+  module HTML4
     class ElementDescription
       ###
       # Is this element a block element?

data/lib/nokogiri/{html → html4}/element_description_defaults.rb

@@ -1,5 +1,6 @@
+# frozen_string_literal: true
 module Nokogiri
-  module HTML
+  module HTML4
     class ElementDescription
 
       # Methods are defined protected by method_defined? because at

data/lib/nokogiri/{html → html4}/entity_lookup.rb

@@ -1,5 +1,6 @@
+# frozen_string_literal: true
 module Nokogiri
-  module HTML
+  module HTML4
     class EntityDescription < Struct.new(:value, :name, :description); end
 
     class EntityLookup