RubyGems - nokogiri - Versions diffs - 1.13.10-arm64-darwin → 1.14.0-arm64-darwin - Mend

nokogiri 1.13.10-arm64-darwin → 1.14.0-arm64-darwin

Potentially problematic release.

This version of nokogiri might be problematic. Click here for more details.

Files changed (69) hide show

checksums.yaml +4 -4
data/Gemfile +33 -0
data/LICENSE-DEPENDENCIES.md +830 -509
data/LICENSE.md +1 -1
data/README.md +18 -11
data/dependencies.yml +25 -7
data/ext/nokogiri/extconf.rb +80 -21
data/ext/nokogiri/gumbo.c +19 -9
data/ext/nokogiri/html4_document.c +1 -1
data/ext/nokogiri/html4_entity_lookup.c +1 -1
data/ext/nokogiri/html4_sax_parser_context.c +0 -5
data/ext/nokogiri/include/libxslt/xsltconfig.h +1 -1
data/ext/nokogiri/nokogiri.c +33 -51
data/ext/nokogiri/nokogiri.h +17 -14
data/ext/nokogiri/xml_attribute_decl.c +1 -1
data/ext/nokogiri/xml_cdata.c +1 -1
data/ext/nokogiri/xml_document.c +16 -11
data/ext/nokogiri/xml_element_content.c +2 -2
data/ext/nokogiri/xml_element_decl.c +1 -1
data/ext/nokogiri/xml_encoding_handler.c +2 -2
data/ext/nokogiri/xml_namespace.c +38 -8
data/ext/nokogiri/xml_node.c +286 -26
data/ext/nokogiri/xml_node_set.c +0 -2
data/ext/nokogiri/xml_reader.c +40 -20
data/ext/nokogiri/xml_relax_ng.c +0 -2
data/ext/nokogiri/xml_sax_parser.c +22 -16
data/ext/nokogiri/xml_sax_parser_context.c +0 -5
data/ext/nokogiri/xml_sax_push_parser.c +0 -2
data/ext/nokogiri/xml_schema.c +0 -2
data/ext/nokogiri/xml_xpath_context.c +87 -83
data/ext/nokogiri/xslt_stylesheet.c +14 -13
data/gumbo-parser/Makefile +10 -0
data/lib/nokogiri/2.7/nokogiri.bundle +0 -0
data/lib/nokogiri/3.0/nokogiri.bundle +0 -0
data/lib/nokogiri/3.1/nokogiri.bundle +0 -0
data/lib/nokogiri/{2.6 → 3.2}/nokogiri.bundle +0 -0
data/lib/nokogiri/css/node.rb +2 -2
data/lib/nokogiri/css/xpath_visitor.rb +5 -3
data/lib/nokogiri/css.rb +6 -0
data/lib/nokogiri/encoding_handler.rb +57 -0
data/lib/nokogiri/extension.rb +3 -2
data/lib/nokogiri/html4/document.rb +2 -121
data/lib/nokogiri/html4/element_description_defaults.rb +6 -12
data/lib/nokogiri/html4/encoding_reader.rb +121 -0
data/lib/nokogiri/html4.rb +1 -0
data/lib/nokogiri/html5/document.rb +113 -36
data/lib/nokogiri/html5/document_fragment.rb +9 -2
data/lib/nokogiri/html5/node.rb +3 -5
data/lib/nokogiri/html5.rb +127 -216
data/lib/nokogiri/jruby/dependencies.rb +1 -19
data/lib/nokogiri/jruby/nokogiri_jars.rb +43 -0
data/lib/nokogiri/version/constant.rb +1 -1
data/lib/nokogiri/version/info.rb +11 -10
data/lib/nokogiri/xml/attr.rb +49 -0
data/lib/nokogiri/xml/builder.rb +1 -1
data/lib/nokogiri/xml/document.rb +102 -54
data/lib/nokogiri/xml/document_fragment.rb +49 -6
data/lib/nokogiri/xml/namespace.rb +42 -0
data/lib/nokogiri/xml/node/save_options.rb +6 -4
data/lib/nokogiri/xml/node.rb +190 -35
data/lib/nokogiri/xml/node_set.rb +87 -9
data/lib/nokogiri/xml/parse_options.rb +129 -50
data/lib/nokogiri/xml/pp/node.rb +6 -4
data/lib/nokogiri/xml/processing_instruction.rb +2 -1
data/lib/nokogiri/xml/sax/parser.rb +2 -3
data/lib/nokogiri/xslt.rb +1 -1
data/lib/nokogiri.rb +3 -11
data/lib/xsd/xmlparser/nokogiri.rb +3 -1
metadata +13 -248

data/lib/nokogiri/xml/node.rb CHANGED Viewed

@@ -137,9 +137,12 @@ module Nokogiri
       ###
       # Add +node_or_tags+ as a child of this Node.
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
       #
-      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is a DocumentFragment, NodeSet, or string).
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # containing markup.
+      #
+      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is
+      # a DocumentFragment, NodeSet, or String).
       #
       # Also see related method +<<+.
       def add_child(node_or_tags)
@@ -154,9 +157,12 @@ module Nokogiri
       ###
       # Add +node_or_tags+ as the first child of this Node.
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
       #
-      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is a DocumentFragment, NodeSet, or string).
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # containing markup.
+      #
+      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is
+      # a DocumentFragment, NodeSet, or String).
       #
       # Also see related method +add_child+.
       def prepend_child(node_or_tags)
@@ -170,22 +176,81 @@ module Nokogiri
         end
       end
-      ###
-      # Add html around this node
+      # :call-seq:
+      #   wrap(markup) -> self
+      #   wrap(node) -> self
+      #
+      # Wrap this Node with the node parsed from +markup+ or a dup of the +node+.
+      #
+      # [Parameters]
+      # - *markup* (String)
+      #   Markup that is parsed and used as the wrapper. This node's parent, if it exists, is used
+      #   as the context node for parsing; otherwise the associated document is used. If the parsed
+      #   fragment has multiple roots, the first root node is used as the wrapper.
+      # - *node* (Nokogiri::XML::Node)
+      #   An element that is `#dup`ed and used as the wrapper.
+      #
+      # [Returns] +self+, to support chaining.
+      #
+      # Also see NodeSet#wrap
+      #
+      # *Example* with a +String+ argument:
       #
-      # Returns self
-      def wrap(html)
-        new_parent = document.parse(html).first
-        add_next_sibling(new_parent)
+      #   doc = Nokogiri::HTML5(<<~HTML)
+      #     <html><body>
+      #       <a>asdf</a>
+      #     </body></html>
+      #   HTML
+      #   doc.at_css("a").wrap("<div></div>")
+      #   doc.to_html
+      #   # => <html><head></head><body>
+      #   #      <div><a>asdf</a></div>
+      #   #    </body></html>
+      #
+      # *Example* with a +Node+ argument:
+      #
+      #   doc = Nokogiri::HTML5(<<~HTML)
+      #     <html><body>
+      #       <a>asdf</a>
+      #     </body></html>
+      #   HTML
+      #   doc.at_css("a").wrap(doc.create_element("div"))
+      #   doc.to_html
+      #   # <html><head></head><body>
+      #   #   <div><a>asdf</a></div>
+      #   # </body></html>
+      #
+      def wrap(node_or_tags)
+        case node_or_tags
+        when String
+          context_node = parent || document
+          new_parent = context_node.coerce(node_or_tags).first
+          if new_parent.nil?
+            raise "Failed to parse '#{node_or_tags}' in the context of a '#{context_node.name}' element"
+          end
+        when XML::Node
+          new_parent = node_or_tags.dup
+        else
+          raise ArgumentError, "Requires a String or Node argument, and cannot accept a #{node_or_tags.class}"
+        end
+        if parent
+          add_next_sibling(new_parent)
+        else
+          new_parent.unlink
+        end
         new_parent.add_child(self)
         self
       end
       ###
       # Add +node_or_tags+ as a child of this Node.
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
       #
-      # Returns self, to support chaining of calls (e.g., root << child1 << child2)
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # containing markup.
+      #
+      # Returns +self+, to support chaining of calls (e.g., root << child1 << child2)
       #
       # Also see related method +add_child+.
       def <<(node_or_tags)
@@ -195,9 +260,12 @@ module Nokogiri
       ###
       # Insert +node_or_tags+ before this Node (as a sibling).
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
       #
-      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is a DocumentFragment, NodeSet, or string).
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # containing markup.
+      #
+      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is
+      # a DocumentFragment, NodeSet, or String).
       #
       # Also see related method +before+.
       def add_previous_sibling(node_or_tags)
@@ -209,9 +277,12 @@ module Nokogiri
       ###
       # Insert +node_or_tags+ after this Node (as a sibling).
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
       #
-      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is a DocumentFragment, NodeSet, or string).
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # containing markup.
+      #
+      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is
+      # a DocumentFragment, NodeSet, or String).
       #
       # Also see related method +after+.
       def add_next_sibling(node_or_tags)
@@ -223,9 +294,11 @@ module Nokogiri
       ####
       # Insert +node_or_tags+ before this node (as a sibling).
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
       #
-      # Returns self, to support chaining of calls.
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # containing markup.
+      #
+      # Returns +self+, to support chaining of calls.
       #
       # Also see related method +add_previous_sibling+.
       def before(node_or_tags)
@@ -235,9 +308,11 @@ module Nokogiri
       ####
       # Insert +node_or_tags+ after this node (as a sibling).
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a Nokogiri::XML::DocumentFragment, or a string containing markup.
       #
-      # Returns self, to support chaining of calls.
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a Nokogiri::XML::DocumentFragment, or a String
+      # containing markup.
+      #
+      # Returns +self+, to support chaining of calls.
       #
       # Also see related method +add_next_sibling+.
       def after(node_or_tags)
@@ -246,8 +321,18 @@ module Nokogiri
       end
       ####
-      # Set the inner html for this Node to +node_or_tags+
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a Nokogiri::XML::DocumentFragment, or a string containing markup.
+      # Set the content for this Node to +node_or_tags+.
+      #
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a Nokogiri::XML::DocumentFragment, or a String
+      # containing markup.
+      #
+      # ⚠ Please note that despite the name, this method will *not* always parse a String argument
+      # as HTML. A String argument will be parsed with the +DocumentFragment+ parser related to this
+      # node's document.
+      #
+      # For example, if the document is an HTML4::Document then the string will be parsed as HTML4
+      # using HTML4::DocumentFragment; but if the document is an XML::Document then it will
+      # parse the string as XML using XML::DocumentFragment.
       #
       # Also see related method +children=+
       def inner_html=(node_or_tags)
@@ -255,8 +340,10 @@ module Nokogiri
       end
       ####
-      # Set the inner html for this Node +node_or_tags+
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a Nokogiri::XML::DocumentFragment, or a string containing markup.
+      # Set the content for this Node +node_or_tags+
+      #
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a Nokogiri::XML::DocumentFragment, or a String
+      # containing markup.
       #
       # Also see related method +inner_html=+
       def children=(node_or_tags)
@@ -271,9 +358,12 @@ module Nokogiri
       ####
       # Replace this Node with +node_or_tags+.
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
       #
-      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is a DocumentFragment, NodeSet, or string).
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # containing markup.
+      #
+      # Returns the reparented node (if +node_or_tags+ is a Node), or NodeSet (if +node_or_tags+ is
+      # a DocumentFragment, NodeSet, or String).
       #
       # Also see related method +swap+.
       def replace(node_or_tags)
@@ -303,7 +393,9 @@ module Nokogiri
       ####
       # Swap this Node for +node_or_tags+
-      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
+      #
+      # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a String
+      # Containing markup.
       #
       # Returns self, to support chaining of calls.
       #
@@ -314,7 +406,8 @@ module Nokogiri
       end
       ####
-      # Set the Node's content to a Text node containing +string+. The string gets XML escaped, not interpreted as markup.
+      # Set the Node's content to a Text node containing +string+. The string gets XML escaped, not
+      # interpreted as markup.
       def content=(string)
         self.native_content = encode_special_chars(string.to_s)
       end
@@ -1105,9 +1198,9 @@ module Nokogiri
       # Get the path to this node as a CSS expression
       def css_path
-        path.split(%r{/}).map do |part|
+        path.split(%r{/}).filter_map do |part|
           part.empty? ? nil : part.gsub(/\[(\d+)\]/, ':nth-of-type(\1)')
-        end.compact.join(" > ")
+        end.join(" > ")
       end
       ###
@@ -1194,12 +1287,11 @@ module Nokogiri
           }
         end
-        encoding = options[:encoding] || document.encoding
-        options[:encoding] = encoding
+        options[:encoding] ||= document.encoding
+        encoding = Encoding.find(options[:encoding] || "UTF-8")
+        io = StringIO.new(String.new(encoding: encoding))
-        outstring = +""
-        outstring.force_encoding(Encoding.find(encoding || "utf-8"))
-        io = StringIO.new(outstring)
         write_to(io, options, &block)
         io.string
       end
@@ -1311,6 +1403,69 @@ module Nokogiri
         end
       end
+      DECONSTRUCT_KEYS = [:name, :attributes, :children, :namespace, :content, :elements, :inner_html].freeze # :nodoc:
+      DECONSTRUCT_METHODS = { attributes: :attribute_nodes }.freeze # :nodoc:
+      #
+      #  :call-seq: deconstruct_keys(array_of_names) → Hash
+      #
+      #  Returns a hash describing the Node, to use in pattern matching.
+      #
+      #  Valid keys and their values:
+      #  - +name+ → (String) The name of this node, or "text" if it is a Text node.
+      #  - +namespace+ → (Namespace, nil) The namespace of this node, or nil if there is no namespace.
+      #  - +attributes+ → (Array<Attr>) The attributes of this node.
+      #  - +children+ → (Array<Node>) The children of this node. 💡 Note this includes text nodes.
+      #  - +elements+ → (Array<Node>) The child elements of this node. 💡 Note this does not include text nodes.
+      #  - +content+ → (String) The contents of all the text nodes in this node's subtree. See #content.
+      #  - +inner_html+ → (String) The inner markup for the children of this node. See #inner_html.
+      #
+      #  ⚡ This is an experimental feature, available since v1.14.0
+      #
+      #  *Example*
+      #
+      #    doc = Nokogiri::XML.parse(<<~XML)
+      #      <?xml version="1.0"?>
+      #      <parent xmlns="http://nokogiri.org/ns/default" xmlns:noko="http://nokogiri.org/ns/noko">
+      #        <child1 foo="abc" noko:bar="def">First</child1>
+      #        <noko:child2 foo="qwe" noko:bar="rty">Second</noko:child2>
+      #      </parent>
+      #    XML
+      #
+      #    doc.root.deconstruct_keys([:name, :namespace])
+      #    # => {:name=>"parent",
+      #    #     :namespace=>
+      #    #      #(Namespace:0x35c { href = "http://nokogiri.org/ns/default" })}
+      #
+      #    doc.root.deconstruct_keys([:inner_html, :content])
+      #    # => {:content=>"\n" + "  First\n" + "  Second\n",
+      #    #     :inner_html=>
+      #    #      "\n" +
+      #    #      "  <child1 foo=\"abc\" noko:bar=\"def\">First</child1>\n" +
+      #    #      "  <noko:child2 foo=\"qwe\" noko:bar=\"rty\">Second</noko:child2>\n"}
+      #
+      #    doc.root.elements.first.deconstruct_keys([:attributes])
+      #    # => {:attributes=>
+      #    #      [#(Attr:0x370 { name = "foo", value = "abc" }),
+      #    #       #(Attr:0x384 {
+      #    #         name = "bar",
+      #    #         namespace = #(Namespace:0x398 {
+      #    #           prefix = "noko",
+      #    #           href = "http://nokogiri.org/ns/noko"
+      #    #           }),
+      #    #         value = "def"
+      #    #         })]}
+      #
+      def deconstruct_keys(keys)
+        requested_keys = DECONSTRUCT_KEYS & keys
+        {}.tap do |values|
+          requested_keys.each do |key|
+            method = DECONSTRUCT_METHODS[key] || key
+            values[key] = send(method)
+          end
+        end
+      end
       # :section:
       protected

data/lib/nokogiri/xml/node_set.rb CHANGED Viewed

@@ -1,3 +1,4 @@
+# coding: utf-8
 # frozen_string_literal: true
 module Nokogiri
@@ -201,7 +202,7 @@ module Nokogiri
       #
       def attr(key, value = nil, &block)
         unless key.is_a?(Hash) || (key && (value || block))
-          return first ? first.attribute(key) : nil
+          return first&.attribute(key)
         end
         hash = key.is_a?(Hash) ? key : { key => value }
@@ -260,10 +261,73 @@ module Nokogiri
         collect { |j| j.inner_html(*args) }.join("")
       end
-      ###
-      # Wrap this NodeSet with +html+
-      def wrap(html)
-        map { |node| node.wrap(html) }
+      # :call-seq:
+      #   wrap(markup) -> self
+      #   wrap(node) -> self
+      #
+      # Wrap each member of this NodeSet with the node parsed from +markup+ or a dup of the +node+.
+      #
+      # [Parameters]
+      # - *markup* (String)
+      #   Markup that is parsed, once per member of the NodeSet, and used as the wrapper. Each
+      #   node's parent, if it exists, is used as the context node for parsing; otherwise the
+      #   associated document is used. If the parsed fragment has multiple roots, the first root
+      #   node is used as the wrapper.
+      # - *node* (Nokogiri::XML::Node)
+      #   An element that is `#dup`ed and used as the wrapper.
+      #
+      # [Returns] +self+, to support chaining.
+      #
+      # ⚠ Note that if a +String+ is passed, the markup will be parsed <b>once per node</b> in the
+      # NodeSet. You can avoid this overhead in cases where you know exactly the wrapper you wish to
+      # use by passing a +Node+ instead.
+      #
+      # Also see Node#wrap
+      #
+      # *Example* with a +String+ argument:
+      #
+      #   doc = Nokogiri::HTML5(<<~HTML)
+      #     <html><body>
+      #       <a>a</a>
+      #       <a>b</a>
+      #       <a>c</a>
+      #       <a>d</a>
+      #     </body></html>
+      #   HTML
+      #   doc.css("a").wrap("<div></div>")
+      #   doc.to_html
+      #   # => <html><head></head><body>
+      #   #      <div><a>a</a></div>
+      #   #      <div><a>b</a></div>
+      #   #      <div><a>c</a></div>
+      #   #      <div><a>d</a></div>
+      #   #    </body></html>
+      #
+      # *Example* with a +Node+ argument
+      #
+      # 💡 Note that this is faster than the equivalent call passing a +String+ because it avoids
+      # having to reparse the wrapper markup for each node.
+      #
+      #   doc = Nokogiri::HTML5(<<~HTML)
+      #     <html><body>
+      #       <a>a</a>
+      #       <a>b</a>
+      #       <a>c</a>
+      #       <a>d</a>
+      #     </body></html>
+      #   HTML
+      #   doc.css("a").wrap(doc.create_element("div"))
+      #   doc.to_html
+      #   # => <html><head></head><body>
+      #   #      <div><a>a</a></div>
+      #   #      <div><a>b</a></div>
+      #   #      <div><a>c</a></div>
+      #   #      <div><a>d</a></div>
+      #   #    </body></html>
+      #
+      def wrap(node_or_tags)
+        map { |node| node.wrap(node_or_tags) }
+        self
       end
       ###
@@ -277,12 +341,15 @@ module Nokogiri
       def to_html(*args)
         if Nokogiri.jruby?
           options = args.first.is_a?(Hash) ? args.shift : {}
-          unless options[:save_with]
-            options[:save_with] = Node::SaveOptions::NO_DECLARATION | Node::SaveOptions::NO_EMPTY_TAGS | Node::SaveOptions::AS_HTML
-          end
+          options[:save_with] ||= Node::SaveOptions::DEFAULT_HTML
           args.insert(0, options)
         end
-        map { |x| x.to_html(*args) }.join
+        if empty?
+          encoding = (args.first.is_a?(Hash) ? args.first[:encoding] : nil) || document.encoding
+          "".encode(encoding)
+        else
+          map { |x| x.to_html(*args) }.join
+        end
       end
       ###
@@ -362,6 +429,17 @@ module Nokogiri
       alias_method :+, :|
+      #
+      #  :call-seq: deconstruct() → Array
+      #
+      #  Returns the members of this NodeSet as an array, to use in pattern matching.
+      #
+      #  ⚡ This is an experimental feature, available since v1.14.0
+      #
+      def deconstruct
+        to_a
+      end
       IMPLIED_XPATH_CONTEXTS = [".//", "self::"].freeze # :nodoc:
     end
   end