nokogiri 1.10.1 → 1.11.0

data/lib/nokogiri/decorators/slop.rb

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

data/lib/nokogiri/html.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 require 'nokogiri/html/entity_lookup'
 require 'nokogiri/html/document'
 require 'nokogiri/html/document_fragment'

data/lib/nokogiri/html/builder.rb

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

data/lib/nokogiri/html/document.rb

@@ -1,3 +1,7 @@
+# frozen_string_literal: true
+
+require 'pathname'
+
 module Nokogiri
   module HTML
     class Document < Nokogiri::XML::Document
@@ -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
 
@@ -272,26 +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

data/lib/nokogiri/html/document_fragment.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 module Nokogiri
   module HTML
     class DocumentFragment < Nokogiri::XML::DocumentFragment

data/lib/nokogiri/html/element_description.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 module Nokogiri
   module HTML
     class ElementDescription

data/lib/nokogiri/html/element_description_defaults.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 module Nokogiri
   module HTML
     class ElementDescription

data/lib/nokogiri/html/entity_lookup.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 module Nokogiri
   module HTML
     class EntityDescription < Struct.new(:value, :name, :description); end

data/lib/nokogiri/html/sax/parser.rb

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

data/lib/nokogiri/html/sax/parser_context.rb

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

data/lib/nokogiri/html/sax/push_parser.rb

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

data/lib/nokogiri/jruby/dependencies.rb

@@ -0,0 +1,20 @@
+# frozen_string_literal: true
+# The line below caused a problem on non-GAE rack environment.
+# unless defined?(JRuby::Rack::VERSION) || defined?(AppEngine::ApiProxy)
+#
+# However, simply cutting defined?(JRuby::Rack::VERSION) off resulted in
+# an unable-to-load-nokogiri problem. Thus, now, Nokogiri checks the presense
+# of appengine-rack.jar in $LOAD_PATH. If Nokogiri is on GAE, Nokogiri
+# should skip loading xml jars. This is because those are in WEB-INF/lib and
+# already set in the classpath.
+unless $LOAD_PATH.to_s.include?("appengine-rack")
+  require 'stringio'
+  require 'isorelax.jar'
+  require 'jing.jar'
+  require 'nekohtml.jar'
+  require 'nekodtd.jar'
+  require 'xercesImpl.jar'
+  require 'serializer.jar'
+  require 'xalan.jar'
+  require 'xml-apis.jar'
+end

data/lib/nokogiri/syntax_error.rb

@@ -1,3 +1,4 @@
+# frozen_string_literal: true
 module Nokogiri
   class SyntaxError < ::StandardError
   end

data/lib/nokogiri/version.rb

@@ -1,109 +1,3 @@
-module Nokogiri
-  # The version of Nokogiri you are using
-  VERSION = '1.10.1'
-
-  class VersionInfo # :nodoc:
-    def jruby?
-      ::JRUBY_VERSION if RUBY_PLATFORM == "java"
-    end
-
-    def engine
-      defined?(RUBY_ENGINE) ? RUBY_ENGINE : 'mri'
-    end
-
-    def loaded_parser_version
-      LIBXML_PARSER_VERSION.
-        scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first.
-        collect(&:to_i).
-        join(".")
-    end
-
-    def compiled_parser_version
-      LIBXML_VERSION
-    end
-
-    def libxml2?
-      defined?(LIBXML_VERSION)
-    end
-
-    def libxml2_using_system?
-      ! libxml2_using_packaged?
-    end
-
-    def libxml2_using_packaged?
-      NOKOGIRI_USE_PACKAGED_LIBRARIES
-    end
-
-    def warnings
-      return [] unless libxml2?
-
-      if compiled_parser_version != loaded_parser_version
-        ["Nokogiri was built against LibXML version #{compiled_parser_version}, but has dynamically loaded #{loaded_parser_version}"]
-      else
-        []
-      end
-    end
-
-    def to_hash
-      hash_info = {}
-      hash_info['warnings']              = []
-      hash_info['nokogiri']              = Nokogiri::VERSION
-      hash_info['ruby']                  = {}
-      hash_info['ruby']['version']       = ::RUBY_VERSION
-      hash_info['ruby']['platform']      = ::RUBY_PLATFORM
-      hash_info['ruby']['description']   = ::RUBY_DESCRIPTION
-      hash_info['ruby']['engine']        = engine
-      hash_info['ruby']['jruby']         = jruby? if jruby?
-
-      if libxml2?
-        hash_info['libxml']              = {}
-        hash_info['libxml']['binding']   = 'extension'
-        if libxml2_using_packaged?
-          hash_info['libxml']['source']  = "packaged"
-          hash_info['libxml']['libxml2_path'] = NOKOGIRI_LIBXML2_PATH
-          hash_info['libxml']['libxslt_path'] = NOKOGIRI_LIBXSLT_PATH
-          hash_info['libxml']['libxml2_patches'] = NOKOGIRI_LIBXML2_PATCHES
-          hash_info['libxml']['libxslt_patches'] = NOKOGIRI_LIBXSLT_PATCHES
-        else
-          hash_info['libxml']['source']  = "system"
-        end
-        hash_info['libxml']['compiled']  = compiled_parser_version
-        hash_info['libxml']['loaded']    = loaded_parser_version
-        hash_info['warnings']            = warnings
-      elsif jruby?
-        hash_info['xerces']   = Nokogiri::XERCES_VERSION
-        hash_info['nekohtml'] = Nokogiri::NEKO_VERSION
-      end
-
-      hash_info
-    end
-
-    def to_markdown
-      begin
-        require 'psych'
-      rescue LoadError
-      end
-      require 'yaml'
-      "# Nokogiri (#{Nokogiri::VERSION})\n" +
-      YAML.dump(to_hash).each_line.map { |line| "    #{line}" }.join
-    end
-
-    # FIXME: maybe switch to singleton?
-    @@instance = new
-    @@instance.warnings.each do |warning|
-      warn "WARNING: #{warning}"
-    end
-    def self.instance; @@instance; end
-  end
-
-  # More complete version information about libxml
-  VERSION_INFO = VersionInfo.instance.to_hash
-
-  def self.uses_libxml? # :nodoc:
-    VersionInfo.instance.libxml2?
-  end
-
-  def self.jruby? # :nodoc:
-    VersionInfo.instance.jruby?
-  end
-end
+# frozen_string_literal: true
+require_relative "version/constant"
+require_relative "version/info"

data/lib/nokogiri/version/constant.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+module Nokogiri
+  # The version of Nokogiri you are using
+  VERSION = "1.11.0"
+end

data/lib/nokogiri/version/info.rb

@@ -0,0 +1,182 @@
+# frozen_string_literal: true
+require "singleton"
+require "shellwords"
+
+module Nokogiri
+  class VersionInfo # :nodoc:
+    include Singleton
+
+    def jruby?
+      ::JRUBY_VERSION if ::RUBY_PLATFORM == "java"
+    end
+
+    def engine
+      defined?(::RUBY_ENGINE) ? ::RUBY_ENGINE : "mri"
+    end
+
+    def loaded_libxml_version
+      Gem::Version.new(Nokogiri::LIBXML_LOADED_VERSION
+        .scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first
+        .collect(&:to_i)
+        .join("."))
+    end
+
+    def compiled_libxml_version
+      Gem::Version.new(Nokogiri::LIBXML_COMPILED_VERSION)
+    end
+
+    def loaded_libxslt_version
+      Gem::Version.new(Nokogiri::LIBXSLT_LOADED_VERSION
+        .scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first
+        .collect(&:to_i)
+        .join("."))
+    end
+
+    def compiled_libxslt_version
+      Gem::Version.new(Nokogiri::LIBXSLT_COMPILED_VERSION)
+    end
+
+    def libxml2?
+      defined?(Nokogiri::LIBXML_COMPILED_VERSION)
+    end
+
+    def libxml2_has_iconv?
+      defined?(Nokogiri::LIBXML_ICONV_ENABLED) && Nokogiri::LIBXML_ICONV_ENABLED
+    end
+
+    def libxml2_using_packaged?
+      libxml2? && Nokogiri::PACKAGED_LIBRARIES
+    end
+
+    def libxml2_using_system?
+      libxml2? && !libxml2_using_packaged?
+    end
+
+    def libxml2_precompiled?
+      libxml2_using_packaged? && Nokogiri::PRECOMPILED_LIBRARIES
+    end
+
+    def warnings
+      warnings = []
+
+      if libxml2?
+        if compiled_libxml_version != loaded_libxml_version
+          warnings << "Nokogiri was built against libxml version #{compiled_libxml_version}, but has dynamically loaded #{loaded_libxml_version}"
+        end
+
+        if compiled_libxslt_version != loaded_libxslt_version
+          warnings << "Nokogiri was built against libxslt version #{compiled_libxslt_version}, but has dynamically loaded #{loaded_libxslt_version}"
+        end
+      end
+
+      warnings
+    end
+
+    def to_hash
+      header_directory = File.expand_path(File.join(File.dirname(__FILE__), "../../../ext/nokogiri"))
+      {}.tap do |vi|
+        vi["warnings"] = []
+        vi["nokogiri"] = {}.tap do |nokogiri|
+          nokogiri["version"] = Nokogiri::VERSION
+
+          unless jruby?
+            cppflags = ["-I#{header_directory.shellescape}"]
+            if libxml2_using_packaged?
+              cppflags << "-I#{File.join(header_directory, "include").shellescape}"
+              cppflags << "-I#{File.join(header_directory, "include/libxml2").shellescape}"
+            end
+            nokogiri["cppflags"] = cppflags
+          end
+        end
+        vi["ruby"] = {}.tap do |ruby|
+          ruby["version"] = ::RUBY_VERSION
+          ruby["platform"] = ::RUBY_PLATFORM
+          ruby["gem_platform"] = ::Gem::Platform.local.to_s
+          ruby["description"] = ::RUBY_DESCRIPTION
+          ruby["engine"] = engine
+          ruby["jruby"] = jruby? if jruby?
+        end
+
+        if libxml2?
+          vi["libxml"] = {}.tap do |libxml|
+            if libxml2_using_packaged?
+              libxml["source"] = "packaged"
+              libxml["precompiled"] = libxml2_precompiled?
+              libxml["patches"] = Nokogiri::LIBXML2_PATCHES
+
+              # this is for nokogumbo and shouldn't be forever
+              libxml["libxml2_path"] = header_directory
+            else
+              libxml["source"] = "system"
+            end
+            libxml["iconv_enabled"] = libxml2_has_iconv?
+            libxml["compiled"] = compiled_libxml_version.to_s
+            libxml["loaded"] = loaded_libxml_version.to_s
+          end
+
+          vi["libxslt"] = {}.tap do |libxslt|
+            if libxml2_using_packaged?
+              libxslt["source"] = "packaged"
+              libxslt["precompiled"] = libxml2_precompiled?
+              libxslt["patches"] = Nokogiri::LIBXSLT_PATCHES
+            else
+              libxslt["source"] = "system"
+            end
+            libxslt["compiled"] = compiled_libxslt_version.to_s
+            libxslt["loaded"] = loaded_libxslt_version.to_s
+          end
+
+          vi["warnings"] = warnings
+        end
+
+        if defined?(Nokogiri::OTHER_LIBRARY_VERSIONS)
+          # see extconf for how this string is assembled: "lib1name:lib1version,lib2name:lib2version"
+          vi["other_libraries"] = Hash[*Nokogiri::OTHER_LIBRARY_VERSIONS.split(/[,:]/)]
+        elsif jruby?
+          vi["other_libraries"] = {}.tap do |ol|
+            ol["xerces"] = Nokogiri::XERCES_VERSION
+            ol["nekohtml"] = Nokogiri::NEKO_VERSION
+          end
+        end
+      end
+    end
+
+    def to_markdown
+      begin
+        require "psych"
+      rescue LoadError
+      end
+      require "yaml"
+      "# Nokogiri (#{Nokogiri::VERSION})\n" +
+      YAML.dump(to_hash).each_line.map { |line| "    #{line}" }.join
+    end
+
+    instance.warnings.each do |warning|
+      warn "WARNING: #{warning}"
+    end
+  end
+
+  def self.uses_libxml?(requirement = nil) # :nodoc:
+    return false unless VersionInfo.instance.libxml2?
+    return true unless requirement
+    Gem::Requirement.new(requirement).satisfied_by?(VersionInfo.instance.loaded_libxml_version)
+  end
+
+  def self.jruby? # :nodoc:
+    VersionInfo.instance.jruby?
+  end
+
+  # Ensure constants used in this file are loaded - see #1896
+  if Nokogiri.jruby?
+    require "nokogiri/jruby/dependencies"
+  end
+  begin
+    ::RUBY_VERSION =~ /(\d+\.\d+)/
+    require "nokogiri/#{Regexp.last_match(1)}/nokogiri"
+  rescue LoadError
+    require "nokogiri/nokogiri"
+  end
+
+  # More complete version information about libxml
+  VERSION_INFO = VersionInfo.instance.to_hash
+end