RubyGems - nokogiri - Versions diffs - 1.14.0.rc1-arm-linux - Mend

nokogiri 1.14.0.rc1-arm-linux

Potentially problematic release.

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

Files changed (200) hide show

checksums.yaml +7 -0
data/Gemfile +38 -0
data/LICENSE-DEPENDENCIES.md +2224 -0
data/LICENSE.md +9 -0
data/README.md +287 -0
data/bin/nokogiri +131 -0
data/dependencies.yml +41 -0
data/ext/nokogiri/depend +38 -0
data/ext/nokogiri/extconf.rb +1082 -0
data/ext/nokogiri/gumbo.c +594 -0
data/ext/nokogiri/html4_document.c +166 -0
data/ext/nokogiri/html4_element_description.c +294 -0
data/ext/nokogiri/html4_entity_lookup.c +37 -0
data/ext/nokogiri/html4_sax_parser_context.c +114 -0
data/ext/nokogiri/html4_sax_push_parser.c +95 -0
data/ext/nokogiri/include/libexslt/exslt.h +108 -0
data/ext/nokogiri/include/libexslt/exsltconfig.h +70 -0
data/ext/nokogiri/include/libexslt/exsltexports.h +63 -0
data/ext/nokogiri/include/libxml2/libxml/HTMLparser.h +306 -0
data/ext/nokogiri/include/libxml2/libxml/HTMLtree.h +147 -0
data/ext/nokogiri/include/libxml2/libxml/SAX.h +204 -0
data/ext/nokogiri/include/libxml2/libxml/SAX2.h +172 -0
data/ext/nokogiri/include/libxml2/libxml/c14n.h +128 -0
data/ext/nokogiri/include/libxml2/libxml/catalog.h +182 -0
data/ext/nokogiri/include/libxml2/libxml/chvalid.h +230 -0
data/ext/nokogiri/include/libxml2/libxml/debugXML.h +217 -0
data/ext/nokogiri/include/libxml2/libxml/dict.h +81 -0
data/ext/nokogiri/include/libxml2/libxml/encoding.h +232 -0
data/ext/nokogiri/include/libxml2/libxml/entities.h +153 -0
data/ext/nokogiri/include/libxml2/libxml/globals.h +499 -0
data/ext/nokogiri/include/libxml2/libxml/hash.h +236 -0
data/ext/nokogiri/include/libxml2/libxml/list.h +137 -0
data/ext/nokogiri/include/libxml2/libxml/nanoftp.h +186 -0
data/ext/nokogiri/include/libxml2/libxml/nanohttp.h +81 -0
data/ext/nokogiri/include/libxml2/libxml/parser.h +1244 -0
data/ext/nokogiri/include/libxml2/libxml/parserInternals.h +656 -0
data/ext/nokogiri/include/libxml2/libxml/pattern.h +100 -0
data/ext/nokogiri/include/libxml2/libxml/relaxng.h +218 -0
data/ext/nokogiri/include/libxml2/libxml/schemasInternals.h +958 -0
data/ext/nokogiri/include/libxml2/libxml/schematron.h +142 -0
data/ext/nokogiri/include/libxml2/libxml/threads.h +91 -0
data/ext/nokogiri/include/libxml2/libxml/tree.h +1312 -0
data/ext/nokogiri/include/libxml2/libxml/uri.h +94 -0
data/ext/nokogiri/include/libxml2/libxml/valid.h +463 -0
data/ext/nokogiri/include/libxml2/libxml/xinclude.h +129 -0
data/ext/nokogiri/include/libxml2/libxml/xlink.h +189 -0
data/ext/nokogiri/include/libxml2/libxml/xmlIO.h +368 -0
data/ext/nokogiri/include/libxml2/libxml/xmlautomata.h +146 -0
data/ext/nokogiri/include/libxml2/libxml/xmlerror.h +947 -0
data/ext/nokogiri/include/libxml2/libxml/xmlexports.h +77 -0
data/ext/nokogiri/include/libxml2/libxml/xmlmemory.h +226 -0
data/ext/nokogiri/include/libxml2/libxml/xmlmodule.h +57 -0
data/ext/nokogiri/include/libxml2/libxml/xmlreader.h +428 -0
data/ext/nokogiri/include/libxml2/libxml/xmlregexp.h +222 -0
data/ext/nokogiri/include/libxml2/libxml/xmlsave.h +88 -0
data/ext/nokogiri/include/libxml2/libxml/xmlschemas.h +246 -0
data/ext/nokogiri/include/libxml2/libxml/xmlschemastypes.h +152 -0
data/ext/nokogiri/include/libxml2/libxml/xmlstring.h +140 -0
data/ext/nokogiri/include/libxml2/libxml/xmlunicode.h +202 -0
data/ext/nokogiri/include/libxml2/libxml/xmlversion.h +503 -0
data/ext/nokogiri/include/libxml2/libxml/xmlwriter.h +488 -0
data/ext/nokogiri/include/libxml2/libxml/xpath.h +575 -0
data/ext/nokogiri/include/libxml2/libxml/xpathInternals.h +632 -0
data/ext/nokogiri/include/libxml2/libxml/xpointer.h +137 -0
data/ext/nokogiri/include/libxslt/attributes.h +38 -0
data/ext/nokogiri/include/libxslt/documents.h +93 -0
data/ext/nokogiri/include/libxslt/extensions.h +262 -0
data/ext/nokogiri/include/libxslt/extra.h +72 -0
data/ext/nokogiri/include/libxslt/functions.h +78 -0
data/ext/nokogiri/include/libxslt/imports.h +75 -0
data/ext/nokogiri/include/libxslt/keys.h +53 -0
data/ext/nokogiri/include/libxslt/namespaces.h +68 -0
data/ext/nokogiri/include/libxslt/numbersInternals.h +73 -0
data/ext/nokogiri/include/libxslt/pattern.h +84 -0
data/ext/nokogiri/include/libxslt/preproc.h +43 -0
data/ext/nokogiri/include/libxslt/security.h +104 -0
data/ext/nokogiri/include/libxslt/templates.h +77 -0
data/ext/nokogiri/include/libxslt/transform.h +207 -0
data/ext/nokogiri/include/libxslt/variables.h +118 -0
data/ext/nokogiri/include/libxslt/xslt.h +110 -0
data/ext/nokogiri/include/libxslt/xsltInternals.h +1982 -0
data/ext/nokogiri/include/libxslt/xsltconfig.h +179 -0
data/ext/nokogiri/include/libxslt/xsltexports.h +64 -0
data/ext/nokogiri/include/libxslt/xsltlocale.h +76 -0
data/ext/nokogiri/include/libxslt/xsltutils.h +310 -0
data/ext/nokogiri/libxml2_backwards_compat.c +121 -0
data/ext/nokogiri/nokogiri.c +259 -0
data/ext/nokogiri/nokogiri.h +235 -0
data/ext/nokogiri/test_global_handlers.c +40 -0
data/ext/nokogiri/xml_attr.c +103 -0
data/ext/nokogiri/xml_attribute_decl.c +70 -0
data/ext/nokogiri/xml_cdata.c +57 -0
data/ext/nokogiri/xml_comment.c +62 -0
data/ext/nokogiri/xml_document.c +689 -0
data/ext/nokogiri/xml_document_fragment.c +44 -0
data/ext/nokogiri/xml_dtd.c +208 -0
data/ext/nokogiri/xml_element_content.c +128 -0
data/ext/nokogiri/xml_element_decl.c +69 -0
data/ext/nokogiri/xml_encoding_handler.c +104 -0
data/ext/nokogiri/xml_entity_decl.c +112 -0
data/ext/nokogiri/xml_entity_reference.c +50 -0
data/ext/nokogiri/xml_namespace.c +186 -0
data/ext/nokogiri/xml_node.c +2425 -0
data/ext/nokogiri/xml_node_set.c +496 -0
data/ext/nokogiri/xml_processing_instruction.c +54 -0
data/ext/nokogiri/xml_reader.c +794 -0
data/ext/nokogiri/xml_relax_ng.c +183 -0
data/ext/nokogiri/xml_sax_parser.c +316 -0
data/ext/nokogiri/xml_sax_parser_context.c +283 -0
data/ext/nokogiri/xml_sax_push_parser.c +166 -0
data/ext/nokogiri/xml_schema.c +282 -0
data/ext/nokogiri/xml_syntax_error.c +85 -0
data/ext/nokogiri/xml_text.c +48 -0
data/ext/nokogiri/xml_xpath_context.c +413 -0
data/ext/nokogiri/xslt_stylesheet.c +363 -0
data/gumbo-parser/CHANGES.md +63 -0
data/gumbo-parser/Makefile +111 -0
data/gumbo-parser/THANKS +27 -0
data/lib/nokogiri/2.7/nokogiri.so +0 -0
data/lib/nokogiri/3.0/nokogiri.so +0 -0
data/lib/nokogiri/3.1/nokogiri.so +0 -0
data/lib/nokogiri/3.2/nokogiri.so +0 -0
data/lib/nokogiri/class_resolver.rb +67 -0
data/lib/nokogiri/css/node.rb +54 -0
data/lib/nokogiri/css/parser.rb +770 -0
data/lib/nokogiri/css/parser.y +277 -0
data/lib/nokogiri/css/parser_extras.rb +96 -0
data/lib/nokogiri/css/syntax_error.rb +9 -0
data/lib/nokogiri/css/tokenizer.rb +155 -0
data/lib/nokogiri/css/tokenizer.rex +56 -0
data/lib/nokogiri/css/xpath_visitor.rb +359 -0
data/lib/nokogiri/css.rb +66 -0
data/lib/nokogiri/decorators/slop.rb +44 -0
data/lib/nokogiri/encoding_handler.rb +57 -0
data/lib/nokogiri/extension.rb +32 -0
data/lib/nokogiri/gumbo.rb +15 -0
data/lib/nokogiri/html.rb +48 -0
data/lib/nokogiri/html4/builder.rb +37 -0
data/lib/nokogiri/html4/document.rb +214 -0
data/lib/nokogiri/html4/document_fragment.rb +54 -0
data/lib/nokogiri/html4/element_description.rb +25 -0
data/lib/nokogiri/html4/element_description_defaults.rb +572 -0
data/lib/nokogiri/html4/encoding_reader.rb +121 -0
data/lib/nokogiri/html4/entity_lookup.rb +15 -0
data/lib/nokogiri/html4/sax/parser.rb +63 -0
data/lib/nokogiri/html4/sax/parser_context.rb +20 -0
data/lib/nokogiri/html4/sax/push_parser.rb +37 -0
data/lib/nokogiri/html4.rb +47 -0
data/lib/nokogiri/html5/document.rb +168 -0
data/lib/nokogiri/html5/document_fragment.rb +90 -0
data/lib/nokogiri/html5/node.rb +98 -0
data/lib/nokogiri/html5.rb +389 -0
data/lib/nokogiri/jruby/dependencies.rb +3 -0
data/lib/nokogiri/jruby/nokogiri_jars.rb +43 -0
data/lib/nokogiri/syntax_error.rb +6 -0
data/lib/nokogiri/version/constant.rb +6 -0
data/lib/nokogiri/version/info.rb +223 -0
data/lib/nokogiri/version.rb +4 -0
data/lib/nokogiri/xml/attr.rb +66 -0
data/lib/nokogiri/xml/attribute_decl.rb +20 -0
data/lib/nokogiri/xml/builder.rb +487 -0
data/lib/nokogiri/xml/cdata.rb +13 -0
data/lib/nokogiri/xml/character_data.rb +9 -0
data/lib/nokogiri/xml/document.rb +471 -0
data/lib/nokogiri/xml/document_fragment.rb +205 -0
data/lib/nokogiri/xml/dtd.rb +34 -0
data/lib/nokogiri/xml/element_content.rb +38 -0
data/lib/nokogiri/xml/element_decl.rb +15 -0
data/lib/nokogiri/xml/entity_decl.rb +21 -0
data/lib/nokogiri/xml/entity_reference.rb +20 -0
data/lib/nokogiri/xml/namespace.rb +58 -0
data/lib/nokogiri/xml/node/save_options.rb +68 -0
data/lib/nokogiri/xml/node.rb +1563 -0
data/lib/nokogiri/xml/node_set.rb +446 -0
data/lib/nokogiri/xml/notation.rb +19 -0
data/lib/nokogiri/xml/parse_options.rb +213 -0
data/lib/nokogiri/xml/pp/character_data.rb +21 -0
data/lib/nokogiri/xml/pp/node.rb +57 -0
data/lib/nokogiri/xml/pp.rb +4 -0
data/lib/nokogiri/xml/processing_instruction.rb +11 -0
data/lib/nokogiri/xml/reader.rb +105 -0
data/lib/nokogiri/xml/relax_ng.rb +38 -0
data/lib/nokogiri/xml/sax/document.rb +167 -0
data/lib/nokogiri/xml/sax/parser.rb +125 -0
data/lib/nokogiri/xml/sax/parser_context.rb +21 -0
data/lib/nokogiri/xml/sax/push_parser.rb +61 -0
data/lib/nokogiri/xml/sax.rb +6 -0
data/lib/nokogiri/xml/schema.rb +73 -0
data/lib/nokogiri/xml/searchable.rb +270 -0
data/lib/nokogiri/xml/syntax_error.rb +72 -0
data/lib/nokogiri/xml/text.rb +11 -0
data/lib/nokogiri/xml/xpath/syntax_error.rb +13 -0
data/lib/nokogiri/xml/xpath.rb +21 -0
data/lib/nokogiri/xml/xpath_context.rb +16 -0
data/lib/nokogiri/xml.rb +76 -0
data/lib/nokogiri/xslt/stylesheet.rb +27 -0
data/lib/nokogiri/xslt.rb +65 -0
data/lib/nokogiri.rb +120 -0
data/lib/xsd/xmlparser/nokogiri.rb +104 -0
metadata +317 -0

data/lib/nokogiri/html4/document.rb ADDED Viewed

@@ -0,0 +1,214 @@
+# coding: utf-8
+# frozen_string_literal: true
+require "pathname"
+module Nokogiri
+  module HTML4
+    class Document < Nokogiri::XML::Document
+      ###
+      # Get the meta tag encoding for this document.  If there is no meta tag,
+      # then nil is returned.
+      def meta_encoding
+        if (meta = at_xpath("//meta[@charset]"))
+          meta[:charset]
+        elsif (meta = meta_content_type)
+          meta["content"][/charset\s*=\s*([\w-]+)/i, 1]
+        end
+      end
+      ###
+      # Set the meta tag encoding for this document.
+      #
+      # If an meta encoding tag is already present, its content is
+      # replaced with the given text.
+      #
+      # Otherwise, this method tries to create one at an appropriate
+      # place supplying head and/or html elements as necessary, which
+      # is inside a head element if any, and before any text node or
+      # content element (typically <body>) if any.
+      #
+      # The result when trying to set an encoding that is different
+      # from the document encoding is undefined.
+      #
+      # Beware in CRuby, that libxml2 automatically inserts a meta tag
+      # into a head element.
+      def meta_encoding=(encoding)
+        if (meta = meta_content_type)
+          meta["content"] = format("text/html; charset=%s", encoding)
+          encoding
+        elsif (meta = at_xpath("//meta[@charset]"))
+          meta["charset"] = encoding
+        else
+          meta = XML::Node.new("meta", self)
+          if (dtd = internal_subset) && dtd.html5_dtd?
+            meta["charset"] = encoding
+          else
+            meta["http-equiv"] = "Content-Type"
+            meta["content"] = format("text/html; charset=%s", encoding)
+          end
+          if (head = at_xpath("//head"))
+            head.prepend_child(meta)
+          else
+            set_metadata_element(meta)
+          end
+          encoding
+        end
+      end
+      def meta_content_type
+        xpath("//meta[@http-equiv and boolean(@content)]").find do |node|
+          node["http-equiv"] =~ /\AContent-Type\z/i
+        end
+      end
+      private :meta_content_type
+      ###
+      # Get the title string of this document.  Return nil if there is
+      # no title tag.
+      def title
+        (title = at_xpath("//title")) && title.inner_text
+      end
+      ###
+      # Set the title string of this document.
+      #
+      # If a title element is already present, its content is replaced
+      # with the given text.
+      #
+      # Otherwise, this method tries to create one at an appropriate
+      # place supplying head and/or html elements as necessary, which
+      # is inside a head element if any, right after a meta
+      # encoding/charset tag if any, and before any text node or
+      # content element (typically <body>) if any.
+      def title=(text)
+        tnode = XML::Text.new(text, self)
+        if (title = at_xpath("//title"))
+          title.children = tnode
+          return text
+        end
+        title = XML::Node.new("title", self) << tnode
+        if (head = at_xpath("//head"))
+          head << title
+        elsif (meta = (at_xpath("//meta[@charset]") || meta_content_type))
+          # better put after charset declaration
+          meta.add_next_sibling(title)
+        else
+          set_metadata_element(title)
+        end
+      end
+      def set_metadata_element(element) # rubocop:disable Naming/AccessorMethodName
+        if (head = at_xpath("//head"))
+          head << element
+        elsif (html = at_xpath("//html"))
+          head = html.prepend_child(XML::Node.new("head", self))
+          head.prepend_child(element)
+        elsif (first = children.find do |node|
+                 case node
+                 when XML::Element, XML::Text
+                   true
+                 end
+               end)
+          # We reach here only if the underlying document model
+          # allows <html>/<head> elements to be omitted and does not
+          # automatically supply them.
+          first.add_previous_sibling(element)
+        else
+          html = add_child(XML::Node.new("html", self))
+          head = html.add_child(XML::Node.new("head", self))
+          head.prepend_child(element)
+        end
+      end
+      private :set_metadata_element
+      ####
+      # Serialize Node using +options+. Save options can also be set using a block.
+      #
+      # See also Nokogiri::XML::Node::SaveOptions and Node@Serialization+and+Generating+Output.
+      #
+      # These two statements are equivalent:
+      #
+      #  node.serialize(:encoding => 'UTF-8', :save_with => FORMAT | AS_XML)
+      #
+      # or
+      #
+      #   node.serialize(:encoding => 'UTF-8') do |config|
+      #     config.format.as_xml
+      #   end
+      #
+      def serialize(options = {})
+        options[:save_with] ||= XML::Node::SaveOptions::DEFAULT_HTML
+        super
+      end
+      ####
+      # Create a Nokogiri::XML::DocumentFragment from +tags+
+      def fragment(tags = nil)
+        DocumentFragment.new(self, tags, root)
+      end
+      # :call-seq:
+      #   xpath_doctype() → Nokogiri::CSS::XPathVisitor::DoctypeConfig
+      #
+      # [Returns] The document type which determines CSS-to-XPath translation.
+      #
+      # See XPathVisitor for more information.
+      def xpath_doctype
+        Nokogiri::CSS::XPathVisitor::DoctypeConfig::HTML4
+      end
+      class << self
+        ###
+        # Parse HTML.  +string_or_io+ may be a String, or any object that
+        # responds to _read_ and _close_ such as an IO, or StringIO.
+        # +url+ is resource where this document is located.  +encoding+ is the
+        # encoding that should be used when processing the document. +options+
+        # is a number that sets options in the parser, such as
+        # 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
+          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 == Encoding::ASCII_8BIT
+              encoding ||= string_or_io.encoding.name
+            end
+          end
+          if string_or_io.respond_to?(:read)
+            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
+              string_or_io = EncodingReader.new(string_or_io)
+              begin
+                return read_io(string_or_io, url, encoding, options.to_i)
+              rescue EncodingReader::EncodingFound => e
+                encoding = e.found_encoding
+              end
+            end
+            return read_io(string_or_io, url, encoding, options.to_i)
+          end
+          # read_memory pukes on empty docs
+          if string_or_io.nil? || string_or_io.empty?
+            return encoding ? new.tap { |i| i.encoding = encoding } : new
+          end
+          encoding ||= EncodingReader.detect_encoding(string_or_io)
+          read_memory(string_or_io, url, encoding, options.to_i)
+        end
+      end
+    end
+  end
+end

data/lib/nokogiri/html4/document_fragment.rb ADDED Viewed

@@ -0,0 +1,54 @@
+# frozen_string_literal: true
+module Nokogiri
+  module HTML4
+    class DocumentFragment < Nokogiri::XML::DocumentFragment
+      ####
+      # Create a Nokogiri::XML::DocumentFragment from +tags+, using +encoding+
+      def self.parse(tags, encoding = nil, options = XML::ParseOptions::DEFAULT_HTML, &block)
+        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
+        doc.encoding = encoding
+        new(doc, tags, nil, options, &block)
+      end
+      def initialize(document, tags = nil, ctx = nil, options = XML::ParseOptions::DEFAULT_HTML)
+        return self unless tags
+        options = Nokogiri::XML::ParseOptions.new(options) if Integer === options
+        yield options if block_given?
+        if ctx
+          preexisting_errors = document.errors.dup
+          node_set = ctx.parse("<div>#{tags}</div>", options)
+          node_set.first.children.each { |child| child.parent = self } unless node_set.empty?
+          self.errors = document.errors - preexisting_errors
+        else
+          # This is a horrible hack, but I don't care
+          path = if /^\s*?<body/i.match?(tags)
+            "/html/body"
+          else
+            "/html/body/node()"
+          end
+          temp_doc = HTML4::Document.parse("<html><body>#{tags}", nil, document.encoding, options)
+          temp_doc.xpath(path).each { |child| child.parent = self }
+          self.errors = temp_doc.errors
+        end
+        children
+      end
+    end
+  end
+end

data/lib/nokogiri/html4/element_description.rb ADDED Viewed

@@ -0,0 +1,25 @@
+# frozen_string_literal: true
+module Nokogiri
+  module HTML4
+    class ElementDescription
+      ###
+      # Is this element a block element?
+      def block?
+        !inline?
+      end
+      ###
+      # Convert this description to a string
+      def to_s
+        "#{name}: #{description}"
+      end
+      ###
+      # Inspection information
+      def inspect
+        "#<#{self.class.name}: #{name} #{description}>"
+      end
+    end
+  end
+end