nokogiri 1.6.0 → 1.13.2

data/lib/nokogiri/xml/schema.rb

@@ -1,11 +1,13 @@
+# frozen_string_literal: true
+
 module Nokogiri
   module XML
     class << self
       ###
       # Create a new Nokogiri::XML::Schema object using a +string_or_io+
       # object.
-      def Schema string_or_io
-        Schema.new(string_or_io)
+      def Schema(string_or_io, options = ParseOptions::DEFAULT_SCHEMA)
+        Schema.new(string_or_io, options)
       end
     end
 
@@ -26,15 +28,23 @@ module Nokogiri
     #   end
     #
     # The list of errors are Nokogiri::XML::SyntaxError objects.
+    #
+    # NOTE: As of v1.11.0, Schema treats inputs as UNTRUSTED by default, and so external entities
+    # are not resolved from the network (`http://` or `ftp://`). Previously, parsing treated
+    # documents as "trusted" by default which was counter to Nokogiri's "untrusted by default"
+    # security policy. If a document is trusted, then the caller may turn off the NONET option via
+    # the ParseOptions to re-enable external entity resolution over a network connection.
     class Schema
       # Errors while parsing the schema file
       attr_accessor :errors
+      # The Nokogiri::XML::ParseOptions used to parse the schema
+      attr_accessor :parse_options
 
       ###
       # Create a new Nokogiri::XML::Schema object using a +string_or_io+
       # object.
-      def self.new string_or_io
-        from_document Nokogiri::XML(string_or_io)
+      def self.new(string_or_io, options = ParseOptions::DEFAULT_SCHEMA)
+        from_document(Nokogiri::XML(string_or_io), options)
       end
 
       ###
@@ -42,9 +52,9 @@ module Nokogiri
       # Nokogiri::XML::Document object, or a filename.  An Array of
       # Nokogiri::XML::SyntaxError objects found while validating the
       # +thing+ is returned.
-      def validate thing
-        if thing.is_a?(Nokogiri::XML::Document) 
-          validate_document(thing) 
+      def validate(thing)
+        if thing.is_a?(Nokogiri::XML::Document)
+          validate_document(thing)
         elsif File.file?(thing)
           validate_file(thing)
         else
@@ -55,8 +65,8 @@ module Nokogiri
       ###
       # Returns true if +thing+ is a valid Nokogiri::XML::Document or
       # file.
-      def valid? thing
-        validate(thing).length == 0
+      def valid?(thing)
+        validate(thing).empty?
       end
     end
   end

data/lib/nokogiri/xml/searchable.rb

@@ -0,0 +1,270 @@
+# coding: utf-8
+# frozen_string_literal: true
+
+module Nokogiri
+  module XML
+    #
+    #  The Searchable module declares the interface used for searching your DOM.
+    #
+    #  It implements the public methods #search, #css, and #xpath,
+    #  as well as allowing specific implementations to specialize some
+    #  of the important behaviors.
+    #
+    module Searchable
+      # Regular expression used by Searchable#search to determine if a query
+      # string is CSS or XPath
+      LOOKS_LIKE_XPATH = %r{^(\./|/|\.\.|\.$)}
+
+      # :section: Searching via XPath or CSS Queries
+
+      ###
+      # call-seq:
+      #   search(*paths, [namespace-bindings, xpath-variable-bindings, custom-handler-class])
+      #
+      # Search this object for +paths+. +paths+ must be one or more XPath or CSS queries:
+      #
+      #   node.search("div.employee", ".//title")
+      #
+      # A hash of namespace bindings may be appended:
+      #
+      #   node.search('.//bike:tire', {'bike' => 'http://schwinn.com/'})
+      #   node.search('bike|tire', {'bike' => 'http://schwinn.com/'})
+      #
+      # For XPath queries, a hash of variable bindings may also be appended to the namespace
+      # bindings. For example:
+      #
+      #   node.search('.//address[@domestic=$value]', nil, {:value => 'Yes'})
+      #
+      # 💡 Custom XPath functions and CSS pseudo-selectors may also be defined. To define custom
+      # functions create a class and implement the function you want to define. The first argument
+      # to the method will be the current matching NodeSet. Any other arguments are ones that you
+      # pass in. Note that this class may appear anywhere in the argument list. For example:
+      #
+      #   handler = Class.new {
+      #     def regex node_set, regex
+      #       node_set.find_all { |node| node['some_attribute'] =~ /#{regex}/ }
+      #     end
+      #   }.new
+      #   node.search('.//title[regex(., "\w+")]', 'div.employee:regex("[0-9]+")', handler)
+      #
+      # See Searchable#xpath and Searchable#css for further usage help.
+      def search(*args)
+        paths, handler, ns, binds = extract_params(args)
+
+        xpaths = paths.map(&:to_s).map do |path|
+          LOOKS_LIKE_XPATH.match?(path) ? path : xpath_query_from_css_rule(path, ns)
+        end.flatten.uniq
+
+        xpath(*(xpaths + [ns, handler, binds].compact))
+      end
+
+      alias_method :/, :search
+
+      ###
+      # call-seq:
+      #   at(*paths, [namespace-bindings, xpath-variable-bindings, custom-handler-class])
+      #
+      # Search this object for +paths+, and return only the first
+      # result. +paths+ must be one or more XPath or CSS queries.
+      #
+      # See Searchable#search for more information.
+      def at(*args)
+        search(*args).first
+      end
+
+      alias_method :%, :at
+
+      ###
+      # call-seq:
+      #   css(*rules, [namespace-bindings, custom-pseudo-class])
+      #
+      # Search this object for CSS +rules+. +rules+ must be one or more CSS
+      # selectors. For example:
+      #
+      #   node.css('title')
+      #   node.css('body h1.bold')
+      #   node.css('div + p.green', 'div#one')
+      #
+      # A hash of namespace bindings may be appended. For example:
+      #
+      #   node.css('bike|tire', {'bike' => 'http://schwinn.com/'})
+      #
+      # 💡 Custom CSS pseudo classes may also be defined which are mapped to a custom XPath
+      # function.  To define custom pseudo classes, create a class and implement the custom pseudo
+      # class you want defined. The first argument to the method will be the matching context
+      # NodeSet. Any other arguments are ones that you pass in. For example:
+      #
+      #   handler = Class.new {
+      #     def regex(node_set, regex)
+      #       node_set.find_all { |node| node['some_attribute'] =~ /#{regex}/ }
+      #     end
+      #   }.new
+      #   node.css('title:regex("\w+")', handler)
+      #
+      # 💡 Some XPath syntax is supported in CSS queries. For example, to query for an attribute:
+      #
+      #   node.css('img > @href') # returns all +href+ attributes on an +img+ element
+      #   node.css('img / @href') # same
+      #
+      #   # ⚠ this returns +class+ attributes from all +div+ elements AND THEIR CHILDREN!
+      #   node.css('div @class')
+      #
+      #   node.css
+      #
+      # 💡 Array-like syntax is supported in CSS queries as an alternative to using +:nth-child()+.
+      #
+      # ⚠ NOTE that indices are 1-based like +:nth-child+ and not 0-based like Ruby Arrays. For
+      # example:
+      #
+      #   # equivalent to 'li:nth-child(2)'
+      #   node.css('li[2]') # retrieve the second li element in a list
+      #
+      # ⚠ NOTE that the CSS query string is case-sensitive with regards to your document type. HTML
+      # tags will match only lowercase CSS queries, so if you search for "H1" in an HTML document,
+      # you'll never find anything. However, "H1" might be found in an XML document, where tags
+      # names are case-sensitive (e.g., "H1" is distinct from "h1").
+      def css(*args)
+        rules, handler, ns, _ = extract_params(args)
+
+        css_internal(self, rules, handler, ns)
+      end
+
+      ##
+      # call-seq:
+      #   at_css(*rules, [namespace-bindings, custom-pseudo-class])
+      #
+      # Search this object for CSS +rules+, and return only the first
+      # match. +rules+ must be one or more CSS selectors.
+      #
+      # See Searchable#css for more information.
+      def at_css(*args)
+        css(*args).first
+      end
+
+      ###
+      # call-seq:
+      #   xpath(*paths, [namespace-bindings, variable-bindings, custom-handler-class])
+      #
+      # Search this node for XPath +paths+. +paths+ must be one or more XPath
+      # queries.
+      #
+      #   node.xpath('.//title')
+      #
+      # A hash of namespace bindings may be appended. For example:
+      #
+      #   node.xpath('.//foo:name', {'foo' => 'http://example.org/'})
+      #   node.xpath('.//xmlns:name', node.root.namespaces)
+      #
+      # A hash of variable bindings may also be appended to the namespace bindings. For example:
+      #
+      #   node.xpath('.//address[@domestic=$value]', nil, {:value => 'Yes'})
+      #
+      # 💡 Custom XPath functions may also be defined. To define custom functions create a class and
+      # implement the function you want to define. The first argument to the method will be the
+      # current matching NodeSet. Any other arguments are ones that you pass in. Note that this
+      # class may appear anywhere in the argument list. For example:
+      #
+      #   handler = Class.new {
+      #     def regex(node_set, regex)
+      #       node_set.find_all { |node| node['some_attribute'] =~ /#{regex}/ }
+      #     end
+      #   }.new
+      #   node.xpath('.//title[regex(., "\w+")]', handler)
+      #
+      def xpath(*args)
+        paths, handler, ns, binds = extract_params(args)
+
+        xpath_internal(self, paths, handler, ns, binds)
+      end
+
+      ##
+      # call-seq:
+      #   at_xpath(*paths, [namespace-bindings, variable-bindings, custom-handler-class])
+      #
+      # Search this node for XPath +paths+, and return only the first
+      # match. +paths+ must be one or more XPath queries.
+      #
+      # See Searchable#xpath for more information.
+      def at_xpath(*args)
+        xpath(*args).first
+      end
+
+      # :call-seq:
+      #   >(selector) → NodeSet
+      #
+      # Search this node's immediate children using CSS selector +selector+
+      def >(selector) # rubocop:disable Naming/BinaryOperatorParameterName
+        ns = (document.root&.namespaces || {})
+        xpath(CSS.xpath_for(selector, prefix: "./", ns: ns).first)
+      end
+
+      # :section:
+
+      private
+
+      def css_internal(node, rules, handler, ns)
+        xpath_internal(node, css_rules_to_xpath(rules, ns), handler, ns, nil)
+      end
+
+      def xpath_internal(node, paths, handler, ns, binds)
+        document = node.document
+        return NodeSet.new(document) unless document
+
+        if paths.length == 1
+          return xpath_impl(node, paths.first, handler, ns, binds)
+        end
+
+        NodeSet.new(document) do |combined|
+          paths.each do |path|
+            xpath_impl(node, path, handler, ns, binds).each { |set| combined << set }
+          end
+        end
+      end
+
+      def xpath_impl(node, path, handler, ns, binds)
+        ctx = XPathContext.new(node)
+        ctx.register_namespaces(ns)
+        path = path.gsub(/xmlns:/, " :") unless Nokogiri.uses_libxml?
+
+        binds&.each do |key, value|
+          ctx.register_variable(key.to_s, value)
+        end
+
+        ctx.evaluate(path, handler)
+      end
+
+      def css_rules_to_xpath(rules, ns)
+        rules.map { |rule| xpath_query_from_css_rule(rule, ns) }
+      end
+
+      def xpath_query_from_css_rule(rule, ns)
+        visitor = Nokogiri::CSS::XPathVisitor.new(
+          builtins: Nokogiri::CSS::XPathVisitor::BuiltinsConfig::OPTIMAL,
+          doctype: document.xpath_doctype,
+        )
+        self.class::IMPLIED_XPATH_CONTEXTS.map do |implied_xpath_context|
+          CSS.xpath_for(rule.to_s, { prefix: implied_xpath_context, ns: ns,
+                                     visitor: visitor, })
+        end.join(" | ")
+      end
+
+      def extract_params(params) # :nodoc:
+        handler = params.find do |param|
+          ![Hash, String, Symbol].include?(param.class)
+        end
+        params -= [handler] if handler
+
+        hashes = []
+        while Hash === params.last || params.last.nil?
+          hashes << params.pop
+          break if params.empty?
+        end
+        ns, binds = hashes.reverse
+
+        ns ||= (document.root&.namespaces || {})
+
+        [params, handler, ns, binds]
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/syntax_error.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module Nokogiri
   module XML
     ###
@@ -40,7 +42,29 @@ module Nokogiri
       end
 
       def to_s
-        super.chomp
+        message = super.chomp
+        [location_to_s, level_to_s, message]
+          .compact.join(": ")
+          .force_encoding(message.encoding)
+      end
+
+      private
+
+      def level_to_s
+        case level
+        when 3 then "FATAL"
+        when 2 then "ERROR"
+        when 1 then "WARNING"
+        end
+      end
+
+      def nil_or_zero?(attribute)
+        attribute.nil? || attribute.zero?
+      end
+
+      def location_to_s
+        return nil if nil_or_zero?(line) && nil_or_zero?(column)
+        "#{line}:#{column}"
       end
     end
   end

data/lib/nokogiri/xml/text.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module Nokogiri
   module XML
     class Text < Nokogiri::XML::CharacterData

data/lib/nokogiri/xml/xpath/syntax_error.rb

@@ -1,9 +1,11 @@
+# frozen_string_literal: true
+
 module Nokogiri
   module XML
-    class XPath
+    module XPath
       class SyntaxError < XML::SyntaxError
         def to_s
-          [super.chomp, str1].compact.join(': ')
+          [super.chomp, str1].compact.join(": ")
         end
       end
     end

data/lib/nokogiri/xml/xpath.rb

@@ -1,10 +1,21 @@
-require 'nokogiri/xml/xpath/syntax_error'
+# frozen_string_literal: true
 
 module Nokogiri
   module XML
-    class XPath
-      # The Nokogiri::XML::Document tied to this XPath instance
-      attr_accessor :document
+    module XPath
+      # The XPath search prefix to search globally, +//+
+      GLOBAL_SEARCH_PREFIX = "//"
+
+      # The XPath search prefix to search direct descendants of the root element, +/+
+      ROOT_SEARCH_PREFIX = "/"
+
+      # The XPath search prefix to search direct descendants of the current element, +./+
+      CURRENT_SEARCH_PREFIX = "./"
+
+      # The XPath search prefix to search anywhere in the current element's subtree, +.//+
+      SUBTREE_SEARCH_PREFIX = ".//"
     end
   end
 end
+
+require_relative "xpath/syntax_error"

data/lib/nokogiri/xml/xpath_context.rb

@@ -1,16 +1,16 @@
+# frozen_string_literal: true
+
 module Nokogiri
   module XML
     class XPathContext
-
       ###
       # Register namespaces in +namespaces+
       def register_namespaces(namespaces)
         namespaces.each do |k, v|
-          k = k.to_s.gsub(/.*:/,'') # strip off 'xmlns:' or 'xml:'
+          k = k.to_s.gsub(/.*:/, "") # strip off 'xmlns:' or 'xml:'
           register_ns(k, v)
         end
       end
-
     end
   end
 end

data/lib/nokogiri/xml.rb

@@ -1,35 +1,10 @@
-require 'nokogiri/xml/pp'
-require 'nokogiri/xml/parse_options'
-require 'nokogiri/xml/sax'
-require 'nokogiri/xml/node'
-require 'nokogiri/xml/attribute_decl'
-require 'nokogiri/xml/element_decl'
-require 'nokogiri/xml/element_content'
-require 'nokogiri/xml/character_data'
-require 'nokogiri/xml/namespace'
-require 'nokogiri/xml/attr'
-require 'nokogiri/xml/dtd'
-require 'nokogiri/xml/cdata'
-require 'nokogiri/xml/text'
-require 'nokogiri/xml/document'
-require 'nokogiri/xml/document_fragment'
-require 'nokogiri/xml/processing_instruction'
-require 'nokogiri/xml/node_set'
-require 'nokogiri/xml/syntax_error'
-require 'nokogiri/xml/xpath'
-require 'nokogiri/xml/xpath_context'
-require 'nokogiri/xml/builder'
-require 'nokogiri/xml/reader'
-require 'nokogiri/xml/notation'
-require 'nokogiri/xml/entity_decl'
-require 'nokogiri/xml/schema'
-require 'nokogiri/xml/relax_ng'
+# frozen_string_literal: true
 
 module Nokogiri
   class << self
     ###
     # Parse XML.  Convenience method for Nokogiri::XML::Document.parse
-    def XML thing, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_XML, &block
+    def XML(thing, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_XML, &block)
       Nokogiri::XML::Document.parse(thing, url, encoding, options, &block)
     end
   end
@@ -38,20 +13,18 @@ module Nokogiri
     # Original C14N 1.0 spec canonicalization
     XML_C14N_1_0 =       0
     # Exclusive C14N 1.0 spec canonicalization
-    XML_C14N_EXCLUSIVE_1_0 =     1
+    XML_C14N_EXCLUSIVE_1_0 = 1
     # C14N 1.1 spec canonicalization
     XML_C14N_1_1 = 2
     class << self
       ###
       # Parse an XML document using the Nokogiri::XML::Reader API.  See
       # Nokogiri::XML::Reader for mor information
-      def Reader string_or_io, url = nil, encoding = nil, options = ParseOptions::STRICT
-
-        options = Nokogiri::XML::ParseOptions.new(options) if Fixnum === options
-        # Give the options to the user
+      def Reader(string_or_io, url = nil, encoding = nil, options = ParseOptions::STRICT)
+        options = Nokogiri::XML::ParseOptions.new(options) if Integer === options
         yield options if block_given?
 
-        if string_or_io.respond_to? :read
+        if string_or_io.respond_to?(:read)
           return Reader.from_io(string_or_io, url, encoding, options.to_i)
         end
         Reader.from_memory(string_or_io, url, encoding, options.to_i)
@@ -59,15 +32,44 @@ module Nokogiri
 
       ###
       # Parse XML.  Convenience method for Nokogiri::XML::Document.parse
-      def parse thing, url = nil, encoding = nil, options = ParseOptions::DEFAULT_XML, &block
+      def parse(thing, url = nil, encoding = nil, options = ParseOptions::DEFAULT_XML, &block)
         Document.parse(thing, url, encoding, options, &block)
       end
 
       ####
       # Parse a fragment from +string+ in to a NodeSet.
-      def fragment string
-        XML::DocumentFragment.parse(string)
+      def fragment(string, options = ParseOptions::DEFAULT_XML, &block)
+        XML::DocumentFragment.parse(string, options, &block)
       end
     end
   end
 end
+
+require_relative "xml/pp"
+require_relative "xml/parse_options"
+require_relative "xml/sax"
+require_relative "xml/searchable"
+require_relative "xml/node"
+require_relative "xml/attribute_decl"
+require_relative "xml/element_decl"
+require_relative "xml/element_content"
+require_relative "xml/character_data"
+require_relative "xml/namespace"
+require_relative "xml/attr"
+require_relative "xml/dtd"
+require_relative "xml/cdata"
+require_relative "xml/text"
+require_relative "xml/document"
+require_relative "xml/document_fragment"
+require_relative "xml/processing_instruction"
+require_relative "xml/node_set"
+require_relative "xml/syntax_error"
+require_relative "xml/xpath"
+require_relative "xml/xpath_context"
+require_relative "xml/builder"
+require_relative "xml/reader"
+require_relative "xml/notation"
+require_relative "xml/entity_decl"
+require_relative "xml/entity_reference"
+require_relative "xml/schema"
+require_relative "xml/relax_ng"

data/lib/nokogiri/xslt/stylesheet.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module Nokogiri
   module XSLT
     ###
@@ -17,7 +19,7 @@ module Nokogiri
       # Apply an XSLT stylesheet to an XML::Document.
       # +params+ is an array of strings used as XSLT parameters.
       # returns serialized document
-      def apply_to document, params = []
+      def apply_to(document, params = [])
         serialize(transform(document, params))
       end
     end

data/lib/nokogiri/xslt.rb

@@ -1,4 +1,5 @@
-require 'nokogiri/xslt/stylesheet'
+# coding: utf-8
+# frozen_string_literal: true
 
 module Nokogiri
   class << self
@@ -9,7 +10,7 @@ module Nokogiri
     #
     #   xslt = Nokogiri::XSLT(File.read(ARGV[0]))
     #
-    def XSLT stylesheet, modules = {}
+    def XSLT(stylesheet, modules = {})
       XSLT.parse(stylesheet, modules)
     end
   end
@@ -21,36 +22,44 @@ module Nokogiri
     class << self
       ###
       # Parse the stylesheet in +string+, register any +modules+
-      def parse string, modules = {}
+      def parse(string, modules = {})
         modules.each do |url, klass|
-          XSLT.register url, klass
+          XSLT.register(url, klass)
         end
 
+        doc = XML::Document.parse(string, nil, nil, XML::ParseOptions::DEFAULT_XSLT)
         if Nokogiri.jruby?
-          Stylesheet.parse_stylesheet_doc(XML.parse(string), string)
+          Stylesheet.parse_stylesheet_doc(doc, string)
         else
-          Stylesheet.parse_stylesheet_doc(XML.parse(string))
+          Stylesheet.parse_stylesheet_doc(doc)
         end
       end
 
-      ###
-      # Quote parameters in +params+ for stylesheet safety
-      def quote_params params
-        parray = (params.instance_of?(Hash) ? params.to_a.flatten : params).dup
-        parray.each_with_index do |v,i|
-          if i % 2 > 0
-            parray[i]=
-              if v =~ /'/
-                "concat('#{ v.gsub(/'/, %q{', "'", '}) }')"
-              else
-                "'#{v}'";
-              end
+      # :call-seq:
+      #   quote_params(params) → Array
+      #
+      # Quote parameters in +params+ for stylesheet safety.
+      # See Nokogiri::XSLT::Stylesheet.transform for example usage.
+      #
+      # [Parameters]
+      # - +params+ (Hash, Array) XSLT parameters (key->value, or tuples of [key, value])
+      #
+      # [Returns] Array of string parameters, with quotes correctly escaped for use with XSLT::Stylesheet.transform
+      #
+      def quote_params(params)
+        params.flatten.each_slice(2).each_with_object([]) do |kv, quoted_params|
+          key, value = kv.map(&:to_s)
+          value = if /'/.match?(value)
+            "concat('#{value.gsub(/'/, %q{', "'", '})}')"
           else
-            parray[i] = v.to_s
+            "'#{value}'"
           end
+          quoted_params << key
+          quoted_params << value
         end
-        parray.flatten
       end
     end
   end
 end
+
+require_relative "xslt/stylesheet"