addressable 2.6.0 → 2.8.5

data/lib/addressable/uri.rb

@@ -1,6 +1,5 @@
 # frozen_string_literal: true
 
-# encoding:utf-8
 #--
 # Copyright (C) Bob Aman
 #
@@ -38,20 +37,35 @@ module Addressable
     ##
     # Container for the character classes specified in
     # <a href="http://www.ietf.org/rfc/rfc3986.txt">RFC 3986</a>.
+    #
+    # Note: Concatenated and interpolated `String`s are not affected by the
+    #       `frozen_string_literal` directive and must be frozen explicitly.
+    #
+    #       Interpolated `String`s *were* frozen this way before Ruby 3.0:
+    #       https://bugs.ruby-lang.org/issues/17104
     module CharacterClasses
       ALPHA = "a-zA-Z"
       DIGIT = "0-9"
       GEN_DELIMS = "\\:\\/\\?\\#\\[\\]\\@"
       SUB_DELIMS = "\\!\\$\\&\\'\\(\\)\\*\\+\\,\\;\\="
-      RESERVED = GEN_DELIMS + SUB_DELIMS
-      UNRESERVED = ALPHA + DIGIT + "\\-\\.\\_\\~"
-      PCHAR = UNRESERVED + SUB_DELIMS + "\\:\\@"
-      SCHEME = ALPHA + DIGIT + "\\-\\+\\."
-      HOST = UNRESERVED + SUB_DELIMS + "\\[\\:\\]"
-      AUTHORITY = PCHAR
-      PATH = PCHAR + "\\/"
-      QUERY = PCHAR + "\\/\\?"
-      FRAGMENT = PCHAR + "\\/\\?"
+      RESERVED = (GEN_DELIMS + SUB_DELIMS).freeze
+      UNRESERVED = (ALPHA + DIGIT + "\\-\\.\\_\\~").freeze
+      PCHAR = (UNRESERVED + SUB_DELIMS + "\\:\\@").freeze
+      SCHEME = (ALPHA + DIGIT + "\\-\\+\\.").freeze
+      HOST = (UNRESERVED + SUB_DELIMS + "\\[\\:\\]").freeze
+      AUTHORITY = (PCHAR + "\\[\\]").freeze
+      PATH = (PCHAR + "\\/").freeze
+      QUERY = (PCHAR + "\\/\\?").freeze
+      FRAGMENT = (PCHAR + "\\/\\?").freeze
+    end
+
+    module NormalizeCharacterClasses
+      HOST = /[^#{CharacterClasses::HOST}]/
+      UNRESERVED = /[^#{CharacterClasses::UNRESERVED}]/
+      PCHAR = /[^#{CharacterClasses::PCHAR}]/
+      SCHEME = /[^#{CharacterClasses::SCHEME}]/
+      FRAGMENT = /[^#{CharacterClasses::FRAGMENT}]/
+      QUERY = %r{[^a-zA-Z0-9\-\.\_\~\!\$\'\(\)\*\+\,\=\:\@\/\?%]|%(?!2B|2b)}
     end
 
     SLASH = '/'
@@ -73,7 +87,7 @@ module Addressable
       "wais" => 210,
       "ldap" => 389,
       "prospero" => 1525
-    }
+    }.freeze
 
     ##
     # Returns a URI object based on the parsed string.
@@ -103,7 +117,7 @@ module Addressable
         uri = uri.to_str
       rescue TypeError, NoMethodError
         raise TypeError, "Can't convert #{uri.class} into String."
-      end if not uri.is_a? String
+      end unless uri.is_a?(String)
 
       # This Regexp supplied as an example in RFC 3986, and it works great.
       scan = uri.scan(URIREGEX)
@@ -124,15 +138,15 @@ module Addressable
           user = userinfo.strip[/^([^:]*):?/, 1]
           password = userinfo.strip[/:(.*)$/, 1]
         end
+
         host = authority.sub(
           /^([^\[\]]*)@/, EMPTY_STR
         ).sub(
           /:([^:@\[\]]*?)$/, EMPTY_STR
         )
+
         port = authority[/:([^:@\[\]]*?)$/, 1]
-      end
-      if port == EMPTY_STR
-        port = nil
+        port = nil if port == EMPTY_STR
       end
 
       return new(
@@ -175,7 +189,7 @@ module Addressable
         uri = uri.to_s
       end
 
-      if !uri.respond_to?(:to_str)
+      unless uri.respond_to?(:to_str)
         raise TypeError, "Can't convert #{uri.class} into String."
       end
       # Otherwise, convert to a String
@@ -207,7 +221,7 @@ module Addressable
       fragments = match.captures
       authority = fragments[3]
       if authority && authority.length > 0
-        new_authority = authority.gsub(/\\/, '/').gsub(/ /, '%20')
+        new_authority = authority.tr("\\", "/").gsub(" ", "%20")
         # NOTE: We want offset 4, not 3!
         offset = match.offset(4)
         uri = uri.dup
@@ -218,8 +232,9 @@ module Addressable
         parsed = self.parse(hints[:scheme] + "://" + uri)
       end
       if parsed.path.include?(".")
-        new_host = parsed.path[/^([^\/]+\.[^\/]*)/, 1]
-        if new_host
+        if parsed.path[/\b@\b/]
+          parsed.scheme = "mailto" unless parsed.scheme
+        elsif new_host = parsed.path[/^([^\/]+\.[^\/]*)/, 1]
           parsed.defer_validation do
             new_path = parsed.path.sub(
               Regexp.new("^" + Regexp.escape(new_host)), EMPTY_STR)
@@ -266,7 +281,7 @@ module Addressable
       return nil unless path
       # If a URI object is passed, just return itself.
       return path if path.kind_of?(self)
-      if !path.respond_to?(:to_str)
+      unless path.respond_to?(:to_str)
         raise TypeError, "Can't convert #{path.class} into String."
       end
       # Otherwise, convert to a String
@@ -281,15 +296,15 @@ module Addressable
         uri.path.sub!(/^\/?([a-zA-Z])[\|:][\\\/]/) do
           "/#{$1.downcase}:/"
         end
-        uri.path.gsub!(/\\/, SLASH)
+        uri.path.tr!("\\", SLASH)
         if File.exist?(uri.path) &&
             File.stat(uri.path).directory?
-          uri.path.sub!(/\/$/, EMPTY_STR)
+          uri.path.chomp!(SLASH)
           uri.path = uri.path + '/'
         end
 
         # If the path is absolute, set the scheme and host.
-        if uri.path =~ /^\//
+        if uri.path.start_with?(SLASH)
           uri.scheme = "file"
           uri.host = EMPTY_STR
         end
@@ -314,18 +329,29 @@ module Addressable
     #   #=> #<Addressable::URI:0xcab390 URI:http://example.com/relative/path>
     def self.join(*uris)
       uri_objects = uris.collect do |uri|
-        if !uri.respond_to?(:to_str)
+        unless uri.respond_to?(:to_str)
           raise TypeError, "Can't convert #{uri.class} into String."
         end
         uri.kind_of?(self) ? uri : self.parse(uri.to_str)
       end
       result = uri_objects.shift.dup
-      for uri in uri_objects
+      uri_objects.each do |uri|
         result.join!(uri)
       end
       return result
     end
 
+    ##
+    # Tables used to optimize encoding operations in `self.encode_component`
+    # and `self.normalize_component`
+    SEQUENCE_ENCODING_TABLE = (0..255).map do |byte|
+      format("%02x", byte).freeze
+    end.freeze
+
+    SEQUENCE_UPCASED_PERCENT_ENCODING_TABLE = (0..255).map do |byte|
+      format("%%%02X", byte).freeze
+    end.freeze
+
     ##
     # Percent encodes a URI component.
     #
@@ -391,19 +417,22 @@ module Addressable
       component = component.dup
       component.force_encoding(Encoding::ASCII_8BIT)
       # Avoiding gsub! because there are edge cases with frozen strings
-      component = component.gsub(character_class) do |sequence|
-        (sequence.unpack('C*').map { |c| "%" + ("%02x" % c).upcase }).join
+      component = component.gsub(character_class) do |char|
+        SEQUENCE_UPCASED_PERCENT_ENCODING_TABLE[char.ord]
       end
       if upcase_encoded.length > 0
-        component = component.gsub(/%(#{upcase_encoded.chars.map do |char|
-          char.unpack('C*').map { |c| '%02x' % c }.join
-        end.join('|')})/i) { |s| s.upcase }
+        upcase_encoded_chars = upcase_encoded.bytes.map do |byte|
+          SEQUENCE_ENCODING_TABLE[byte]
+        end
+        component = component.gsub(/%(#{upcase_encoded_chars.join('|')})/,
+                                   &:upcase)
       end
+
       return component
     end
 
     class << self
-      alias_method :encode_component, :encode_component
+      alias_method :escape_component, :encode_component
     end
 
     ##
@@ -442,16 +471,14 @@ module Addressable
           "Expected Class (String or Addressable::URI), " +
           "got #{return_type.inspect}"
       end
-      uri = uri.dup
-      # Seriously, only use UTF-8. I'm really not kidding!
-      uri.force_encoding("utf-8")
-      leave_encoded = leave_encoded.dup.force_encoding("utf-8")
-      result = uri.gsub(/%[0-9a-f]{2}/iu) do |sequence|
+
+      result = uri.gsub(/%[0-9a-f]{2}/i) do |sequence|
         c = sequence[1..3].to_i(16).chr
-        c.force_encoding("utf-8")
+        c.force_encoding(sequence.encoding)
         leave_encoded.include?(c) ? sequence : c
       end
-      result.force_encoding("utf-8")
+
+      result.force_encoding(Encoding::UTF_8)
       if return_type == String
         return result
       elsif return_type == ::Addressable::URI
@@ -530,13 +557,16 @@ module Addressable
         leave_re = if leave_encoded.length > 0
           character_class = "#{character_class}%" unless character_class.include?('%')
 
-          "|%(?!#{leave_encoded.chars.map do |char|
-            seq = char.unpack('C*').map { |c| '%02x' % c }.join
-            [seq.upcase, seq.downcase]
-          end.flatten.join('|')})"
+          bytes = leave_encoded.bytes
+          leave_encoded_pattern = bytes.map { |b| SEQUENCE_ENCODING_TABLE[b] }.join('|')
+          "|%(?!#{leave_encoded_pattern}|#{leave_encoded_pattern.upcase})"
         end
 
-        character_class = /[^#{character_class}]#{leave_re}/
+        character_class = if leave_re
+                            /[^#{character_class}]#{leave_re}/
+                          else
+                            /[^#{character_class}]/
+                          end
       end
       # We can't perform regexps on invalid UTF sequences, but
       # here we need to, so switch to ASCII.
@@ -545,7 +575,7 @@ module Addressable
       unencoded = self.unencode_component(component, String, leave_encoded)
       begin
         encoded = self.encode_component(
-          Addressable::IDNA.unicode_normalize_kc(unencoded),
+          unencoded.unicode_normalize(:nfc),
           character_class,
           leave_encoded
         )
@@ -653,8 +683,7 @@ module Addressable
       components.each do |key, value|
         if value != nil
           begin
-            components[key] =
-              Addressable::IDNA.unicode_normalize_kc(value.to_str)
+            components[key] = value.to_str.unicode_normalize(:nfc)
           rescue ArgumentError
             # Likely a malformed UTF-8 character, skip unicode normalization
             components[key] = value.to_str
@@ -802,7 +831,9 @@ module Addressable
         end
       end
 
-      self.defer_validation do
+      reset_ivs
+
+      defer_validation do
         # Bunch of crazy logic required because of the composite components
         # like userinfo and authority.
         self.scheme = options[:scheme] if options[:scheme]
@@ -817,7 +848,8 @@ module Addressable
         self.query_values = options[:query_values] if options[:query_values]
         self.fragment = options[:fragment] if options[:fragment]
       end
-      self.to_s
+
+      to_s # force path validation
     end
 
     ##
@@ -844,9 +876,7 @@ module Addressable
     # The scheme component for this URI.
     #
     # @return [String] The scheme component.
-    def scheme
-      return defined?(@scheme) ? @scheme : nil
-    end
+    attr_reader :scheme
 
     ##
     # The scheme component for this URI, normalized.
@@ -854,18 +884,18 @@ module Addressable
     # @return [String] The scheme component, normalized.
     def normalized_scheme
       return nil unless self.scheme
-      @normalized_scheme ||= begin
-        if self.scheme =~ /^\s*ssh\+svn\s*$/i
+      if @normalized_scheme == NONE
+        @normalized_scheme = if self.scheme =~ /^\s*ssh\+svn\s*$/i
           "svn+ssh".dup
         else
           Addressable::URI.normalize_component(
             self.scheme.strip.downcase,
-            Addressable::URI::CharacterClasses::SCHEME
+            Addressable::URI::NormalizeCharacterClasses::SCHEME
           )
         end
       end
       # All normalized values should be UTF-8
-      @normalized_scheme.force_encoding(Encoding::UTF_8) if @normalized_scheme
+      force_utf8_encoding_if_needed(@normalized_scheme)
       @normalized_scheme
     end
 
@@ -880,13 +910,13 @@ module Addressable
         new_scheme = new_scheme.to_str
       end
       if new_scheme && new_scheme !~ /\A[a-z][a-z0-9\.\+\-]*\z/i
-        raise InvalidURIError, "Invalid scheme format: #{new_scheme}"
+        raise InvalidURIError, "Invalid scheme format: '#{new_scheme}'"
       end
       @scheme = new_scheme
       @scheme = nil if @scheme.to_s.strip.empty?
 
       # Reset dependent values
-      remove_instance_variable(:@normalized_scheme) if defined?(@normalized_scheme)
+      @normalized_scheme = NONE
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -897,9 +927,7 @@ module Addressable
     # The user component for this URI.
     #
     # @return [String] The user component.
-    def user
-      return defined?(@user) ? @user : nil
-    end
+    attr_reader :user
 
     ##
     # The user component for this URI, normalized.
@@ -907,20 +935,20 @@ module Addressable
     # @return [String] The user component, normalized.
     def normalized_user
       return nil unless self.user
-      return @normalized_user if defined?(@normalized_user)
-      @normalized_user ||= begin
+      return @normalized_user unless @normalized_user == NONE
+      @normalized_user = begin
         if normalized_scheme =~ /https?/ && self.user.strip.empty? &&
             (!self.password || self.password.strip.empty?)
           nil
         else
           Addressable::URI.normalize_component(
             self.user.strip,
-            Addressable::URI::CharacterClasses::UNRESERVED
+            Addressable::URI::NormalizeCharacterClasses::UNRESERVED
           )
         end
       end
       # All normalized values should be UTF-8
-      @normalized_user.force_encoding(Encoding::UTF_8) if @normalized_user
+      force_utf8_encoding_if_needed(@normalized_user)
       @normalized_user
     end
 
@@ -936,14 +964,14 @@ module Addressable
 
       # You can't have a nil user with a non-nil password
       if password != nil
-        @user = EMPTY_STR if @user.nil?
+        @user = EMPTY_STR unless user
       end
 
       # Reset dependent values
-      remove_instance_variable(:@userinfo) if defined?(@userinfo)
-      remove_instance_variable(:@normalized_userinfo) if defined?(@normalized_userinfo)
-      remove_instance_variable(:@authority) if defined?(@authority)
-      remove_instance_variable(:@normalized_user) if defined?(@normalized_user)
+      @userinfo = nil
+      @normalized_userinfo = NONE
+      @authority = nil
+      @normalized_user = NONE
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -954,9 +982,7 @@ module Addressable
     # The password component for this URI.
     #
     # @return [String] The password component.
-    def password
-      return defined?(@password) ? @password : nil
-    end
+    attr_reader :password
 
     ##
     # The password component for this URI, normalized.
@@ -964,22 +990,20 @@ module Addressable
     # @return [String] The password component, normalized.
     def normalized_password
       return nil unless self.password
-      return @normalized_password if defined?(@normalized_password)
-      @normalized_password ||= begin
+      return @normalized_password unless @normalized_password == NONE
+      @normalized_password = begin
         if self.normalized_scheme =~ /https?/ && self.password.strip.empty? &&
             (!self.user || self.user.strip.empty?)
           nil
         else
           Addressable::URI.normalize_component(
             self.password.strip,
-            Addressable::URI::CharacterClasses::UNRESERVED
+            Addressable::URI::NormalizeCharacterClasses::UNRESERVED
           )
         end
       end
       # All normalized values should be UTF-8
-      if @normalized_password
-        @normalized_password.force_encoding(Encoding::UTF_8)
-      end
+      force_utf8_encoding_if_needed(@normalized_password)
       @normalized_password
     end
 
@@ -994,17 +1018,15 @@ module Addressable
       @password = new_password ? new_password.to_str : nil
 
       # You can't have a nil user with a non-nil password
-      @password ||= nil
-      @user ||= nil
       if @password != nil
-        @user = EMPTY_STR if @user.nil?
+        self.user = EMPTY_STR if user.nil?
       end
 
       # Reset dependent values
-      remove_instance_variable(:@userinfo) if defined?(@userinfo)
-      remove_instance_variable(:@normalized_userinfo) if defined?(@normalized_userinfo)
-      remove_instance_variable(:@authority) if defined?(@authority)
-      remove_instance_variable(:@normalized_password) if defined?(@normalized_password)
+      @userinfo = nil
+      @normalized_userinfo = NONE
+      @authority = nil
+      @normalized_password = NONE
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1034,8 +1056,8 @@ module Addressable
     # @return [String] The userinfo component, normalized.
     def normalized_userinfo
       return nil unless self.userinfo
-      return @normalized_userinfo if defined?(@normalized_userinfo)
-      @normalized_userinfo ||= begin
+      return @normalized_userinfo unless @normalized_userinfo == NONE
+      @normalized_userinfo = begin
         current_user = self.normalized_user
         current_password = self.normalized_password
         if !current_user && !current_password
@@ -1047,9 +1069,7 @@ module Addressable
         end
       end
       # All normalized values should be UTF-8
-      if @normalized_userinfo
-        @normalized_userinfo.force_encoding(Encoding::UTF_8)
-      end
+      force_utf8_encoding_if_needed(@normalized_userinfo)
       @normalized_userinfo
     end
 
@@ -1075,7 +1095,7 @@ module Addressable
       self.user = new_user
 
       # Reset dependent values
-      remove_instance_variable(:@authority) if defined?(@authority)
+      @authority = nil
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1086,9 +1106,7 @@ module Addressable
     # The host component for this URI.
     #
     # @return [String] The host component.
-    def host
-      return defined?(@host) ? @host : nil
-    end
+    attr_reader :host
 
     ##
     # The host component for this URI, normalized.
@@ -1096,6 +1114,7 @@ module Addressable
     # @return [String] The host component, normalized.
     def normalized_host
       return nil unless self.host
+
       @normalized_host ||= begin
         if !self.host.strip.empty?
           result = ::Addressable::IDNA.to_ascii(
@@ -1107,14 +1126,15 @@ module Addressable
           end
           result = Addressable::URI.normalize_component(
             result,
-            CharacterClasses::HOST)
+            NormalizeCharacterClasses::HOST
+          )
           result
         else
           EMPTY_STR.dup
         end
       end
       # All normalized values should be UTF-8
-      @normalized_host.force_encoding(Encoding::UTF_8) if @normalized_host
+      force_utf8_encoding_if_needed(@normalized_host)
       @normalized_host
     end
 
@@ -1129,8 +1149,8 @@ module Addressable
       @host = new_host ? new_host.to_str : nil
 
       # Reset dependent values
-      remove_instance_variable(:@authority) if defined?(@authority)
-      remove_instance_variable(:@normalized_host) if defined?(@normalized_host)
+      @authority = nil
+      @normalized_host = nil
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1172,7 +1192,7 @@ module Addressable
     # Returns the top-level domain for this host.
     #
     # @example
-    #   Addressable::URI.parse("www.example.co.uk").tld # => "co.uk"
+    #   Addressable::URI.parse("http://www.example.co.uk").tld # => "co.uk"
     def tld
       PublicSuffix.parse(self.host, ignore_private: true).tld
     end
@@ -1182,7 +1202,7 @@ module Addressable
     #
     # @param [String, #to_str] new_tld The new top-level domain.
     def tld=(new_tld)
-      replaced_tld = domain.sub(/#{tld}\z/, new_tld)
+      replaced_tld = host.sub(/#{tld}\z/, new_tld)
       self.host = PublicSuffix::Domain.new(replaced_tld).to_s
     end
 
@@ -1190,7 +1210,7 @@ module Addressable
     # Returns the public suffix domain for this host.
     #
     # @example
-    #   Addressable::URI.parse("www.example.co.uk").domain # => "example.co.uk"
+    #   Addressable::URI.parse("http://www.example.co.uk").domain # => "example.co.uk"
     def domain
       PublicSuffix.domain(self.host, ignore_private: true)
     end
@@ -1232,9 +1252,7 @@ module Addressable
         authority
       end
       # All normalized values should be UTF-8
-      if @normalized_authority
-        @normalized_authority.force_encoding(Encoding::UTF_8)
-      end
+      force_utf8_encoding_if_needed(@normalized_authority)
       @normalized_authority
     end
 
@@ -1263,14 +1281,14 @@ module Addressable
       end
 
       # Password assigned first to ensure validity in case of nil
-      self.password = defined?(new_password) ? new_password : nil
-      self.user = defined?(new_user) ? new_user : nil
-      self.host = defined?(new_host) ? new_host : nil
-      self.port = defined?(new_port) ? new_port : nil
+      self.password = new_password
+      self.user = new_user
+      self.host = new_host
+      self.port = new_port
 
       # Reset dependent values
-      remove_instance_variable(:@userinfo) if defined?(@userinfo)
-      remove_instance_variable(:@normalized_userinfo) if defined?(@normalized_userinfo)
+      @userinfo = nil
+      @normalized_userinfo = NONE
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1318,16 +1336,16 @@ module Addressable
         new_port = new_origin[/:([^:@\[\]\/]*?)$/, 1]
       end
 
-      self.scheme = defined?(new_scheme) ? new_scheme : nil
-      self.host = defined?(new_host) ? new_host : nil
-      self.port = defined?(new_port) ? new_port : nil
+      self.scheme = new_scheme
+      self.host = new_host
+      self.port = new_port
       self.userinfo = nil
 
       # Reset dependent values
-      remove_instance_variable(:@userinfo) if defined?(@userinfo)
-      remove_instance_variable(:@normalized_userinfo) if defined?(@normalized_userinfo)
-      remove_instance_variable(:@authority) if defined?(@authority)
-      remove_instance_variable(:@normalized_authority) if defined?(@normalized_authority)
+      @userinfo = nil
+      @normalized_userinfo = NONE
+      @authority = nil
+      @normalized_authority = nil
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1354,9 +1372,7 @@ module Addressable
     # infer port numbers from default values.
     #
     # @return [Integer] The port component.
-    def port
-      return defined?(@port) ? @port : nil
-    end
+    attr_reader :port
 
     ##
     # The port component for this URI, normalized.
@@ -1364,8 +1380,8 @@ module Addressable
     # @return [Integer] The port component, normalized.
     def normalized_port
       return nil unless self.port
-      return @normalized_port if defined?(@normalized_port)
-      @normalized_port ||= begin
+      return @normalized_port unless @normalized_port == NONE
+      @normalized_port = begin
         if URI.port_mapping[self.normalized_scheme] == self.port
           nil
         else
@@ -1396,8 +1412,8 @@ module Addressable
       @port = nil if @port == 0
 
       # Reset dependent values
-      remove_instance_variable(:@authority) if defined?(@authority)
-      remove_instance_variable(:@normalized_port) if defined?(@normalized_port)
+      @authority = nil
+      @normalized_port = NONE
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1468,7 +1484,7 @@ module Addressable
         site_string
       end
       # All normalized values should be UTF-8
-      @normalized_site.force_encoding(Encoding::UTF_8) if @normalized_site
+      force_utf8_encoding_if_needed(@normalized_site)
       @normalized_site
     end
 
@@ -1498,9 +1514,7 @@ module Addressable
     # The path component for this URI.
     #
     # @return [String] The path component.
-    def path
-      return defined?(@path) ? @path : EMPTY_STR
-    end
+    attr_reader :path
 
     NORMPATH = /^(?!\/)[^\/:]*:.*$/
     ##
@@ -1519,7 +1533,7 @@ module Addressable
         result = path.strip.split(SLASH, -1).map do |segment|
           Addressable::URI.normalize_component(
             segment,
-            Addressable::URI::CharacterClasses::PCHAR
+            Addressable::URI::NormalizeCharacterClasses::PCHAR
           )
         end.join(SLASH)
 
@@ -1531,7 +1545,7 @@ module Addressable
         result
       end
       # All normalized values should be UTF-8
-      @normalized_path.force_encoding(Encoding::UTF_8) if @normalized_path
+      force_utf8_encoding_if_needed(@normalized_path)
       @normalized_path
     end
 
@@ -1549,7 +1563,7 @@ module Addressable
       end
 
       # Reset dependent values
-      remove_instance_variable(:@normalized_path) if defined?(@normalized_path)
+      @normalized_path = nil
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1579,9 +1593,7 @@ module Addressable
     # The query component for this URI.
     #
     # @return [String] The query component.
-    def query
-      return defined?(@query) ? @query : nil
-    end
+    attr_reader :query
 
     ##
     # The query component for this URI, normalized.
@@ -1589,20 +1601,25 @@ module Addressable
     # @return [String] The query component, normalized.
     def normalized_query(*flags)
       return nil unless self.query
-      return @normalized_query if defined?(@normalized_query)
-      @normalized_query ||= begin
+      return @normalized_query unless @normalized_query == NONE
+      @normalized_query = begin
         modified_query_class = Addressable::URI::CharacterClasses::QUERY.dup
         # Make sure possible key-value pair delimiters are escaped.
         modified_query_class.sub!("\\&", "").sub!("\\;", "")
-        pairs = (self.query || "").split("&", -1)
+        pairs = (query || "").split("&", -1)
+        pairs.delete_if(&:empty?).uniq! if flags.include?(:compacted)
         pairs.sort! if flags.include?(:sorted)
         component = pairs.map do |pair|
-          Addressable::URI.normalize_component(pair, modified_query_class, "+")
+          Addressable::URI.normalize_component(
+            pair,
+            Addressable::URI::NormalizeCharacterClasses::QUERY,
+            "+"
+          )
         end.join("&")
         component == "" ? nil : component
       end
       # All normalized values should be UTF-8
-      @normalized_query.force_encoding(Encoding::UTF_8) if @normalized_query
+      force_utf8_encoding_if_needed(@normalized_query)
       @normalized_query
     end
 
@@ -1617,7 +1634,7 @@ module Addressable
       @query = new_query ? new_query.to_str : nil
 
       # Reset dependent values
-      remove_instance_variable(:@normalized_query) if defined?(@normalized_query)
+      @normalized_query = NONE
       remove_composite_values
     end
 
@@ -1656,11 +1673,13 @@ module Addressable
         # so it's best to make all changes in-place.
         pair[0] = URI.unencode_component(pair[0])
         if pair[1].respond_to?(:to_str)
+          value = pair[1].to_str
           # I loathe the fact that I have to do this. Stupid HTML 4.01.
           # Treating '+' as a space was just an unbelievably bad idea.
           # There was nothing wrong with '%20'!
           # If it ain't broke, don't fix it!
-          pair[1] = URI.unencode_component(pair[1].to_str.gsub(/\+/, " "))
+          value = value.tr("+", " ") if ["http", "https", nil].include?(scheme)
+          pair[1] = URI.unencode_component(value)
         end
         if return_type == Hash
           accu[pair[0]] = pair[1]
@@ -1777,9 +1796,7 @@ module Addressable
     # The fragment component for this URI.
     #
     # @return [String] The fragment component.
-    def fragment
-      return defined?(@fragment) ? @fragment : nil
-    end
+    attr_reader :fragment
 
     ##
     # The fragment component for this URI, normalized.
@@ -1787,18 +1804,16 @@ module Addressable
     # @return [String] The fragment component, normalized.
     def normalized_fragment
       return nil unless self.fragment
-      return @normalized_fragment if defined?(@normalized_fragment)
-      @normalized_fragment ||= begin
+      return @normalized_fragment unless @normalized_fragment == NONE
+      @normalized_fragment = begin
         component = Addressable::URI.normalize_component(
           self.fragment,
-          Addressable::URI::CharacterClasses::FRAGMENT
+          Addressable::URI::NormalizeCharacterClasses::FRAGMENT
         )
         component == "" ? nil : component
       end
       # All normalized values should be UTF-8
-      if @normalized_fragment
-        @normalized_fragment.force_encoding(Encoding::UTF_8)
-      end
+      force_utf8_encoding_if_needed(@normalized_fragment)
       @normalized_fragment
     end
 
@@ -1813,7 +1828,7 @@ module Addressable
       @fragment = new_fragment ? new_fragment.to_str : nil
 
       # Reset dependent values
-      remove_instance_variable(:@normalized_fragment) if defined?(@normalized_fragment)
+      @normalized_fragment = NONE
       remove_composite_values
 
       # Ensure we haven't created an invalid URI
@@ -1917,7 +1932,7 @@ module Addressable
               # Section 5.2.3 of RFC 3986
               #
               # Removes the right-most path segment from the base path.
-              if base_path =~ /\//
+              if base_path.include?(SLASH)
                 base_path.sub!(/\/[^\/]+$/, SLASH)
               else
                 base_path = EMPTY_STR
@@ -1979,7 +1994,7 @@ module Addressable
     #
     # @see Hash#merge
     def merge(hash)
-      if !hash.respond_to?(:to_hash)
+      unless hash.respond_to?(:to_hash)
         raise TypeError, "Can't convert #{hash.class} into Hash."
       end
       hash = hash.to_hash
@@ -2367,13 +2382,33 @@ module Addressable
     #
     # @param [Proc] block
     #   A set of operations to perform on a given URI.
-    def defer_validation(&block)
-      raise LocalJumpError, "No block given." unless block
+    def defer_validation
+      raise LocalJumpError, "No block given." unless block_given?
       @validation_deferred = true
-      block.call()
+      yield
       @validation_deferred = false
       validate
-      return nil
+    ensure
+      @validation_deferred = false
+    end
+
+    def encode_with(coder)
+      instance_variables.each do |ivar|
+        value = instance_variable_get(ivar)
+        if value != NONE
+          key = ivar.to_s.slice(1..-1)
+          coder[key] = value
+        end
+      end
+      nil
+    end
+
+    def init_with(coder)
+      reset_ivs
+      coder.map.each do |key, value|
+        instance_variable_set("@#{key}", value)
+      end
+      nil
     end
 
   protected
@@ -2394,30 +2429,35 @@ module Addressable
     def self.normalize_path(path)
       # Section 5.2.4 of RFC 3986
 
-      return nil if path.nil?
+      return if path.nil?
       normalized_path = path.dup
-      begin
-        mod = nil
+      loop do
         mod ||= normalized_path.gsub!(RULE_2A, SLASH)
 
         pair = normalized_path.match(RULE_2B_2C)
-        parent, current = pair[1], pair[2] if pair
+        if pair
+          parent  = pair[1]
+          current = pair[2]
+        else
+          parent  = nil
+          current = nil
+        end
+
+        regexp = "/#{Regexp.escape(parent.to_s)}/\\.\\./|"
+        regexp += "(/#{Regexp.escape(current.to_s)}/\\.\\.$)"
+
         if pair && ((parent != SELF_REF && parent != PARENT) ||
             (current != SELF_REF && current != PARENT))
-          mod ||= normalized_path.gsub!(
-            Regexp.new(
-              "/#{Regexp.escape(parent.to_s)}/\\.\\./|" +
-              "(/#{Regexp.escape(current.to_s)}/\\.\\.$)"
-            ), SLASH
-          )
+          mod ||= normalized_path.gsub!(Regexp.new(regexp), SLASH)
         end
 
         mod ||= normalized_path.gsub!(RULE_2D, EMPTY_STR)
         # Non-standard, removes prefixed dotted segments from path.
         mod ||= normalized_path.gsub!(RULE_PREFIXED_PARENT, SLASH)
-      end until mod.nil?
+        break if mod.nil?
+      end
 
-      return normalized_path
+      normalized_path
     end
 
     ##
@@ -2467,11 +2507,7 @@ module Addressable
     # @return [Addressable::URI] <code>self</code>.
     def replace_self(uri)
       # Reset dependent values
-      instance_variables.each do |var|
-        if instance_variable_defined?(var) && var != :@validation_deferred
-          remove_instance_variable(var)
-        end
-      end
+      reset_ivs
 
       @scheme = uri.scheme
       @user = uri.user
@@ -2503,8 +2539,53 @@ module Addressable
     #
     # @api private
     def remove_composite_values
-      remove_instance_variable(:@uri_string) if defined?(@uri_string)
-      remove_instance_variable(:@hash) if defined?(@hash)
+      @uri_string = nil
+      @hash = nil
     end
+
+    ##
+    # Converts the string to be UTF-8 if it is not already UTF-8
+    #
+    # @api private
+    def force_utf8_encoding_if_needed(str)
+      if str && str.encoding != Encoding::UTF_8
+        str.force_encoding(Encoding::UTF_8)
+      end
+    end
+
+    private
+
+    ##
+    # Resets instance variables
+    #
+    # @api private
+    def reset_ivs
+      @scheme = nil
+      @user = nil
+      @normalized_scheme = NONE
+      @normalized_user = NONE
+      @uri_string = nil
+      @hash = nil
+      @userinfo = nil
+      @normalized_userinfo = NONE
+      @authority = nil
+      @password = nil
+      @normalized_authority = nil
+      @port = nil
+      @normalized_password = NONE
+      @host = nil
+      @normalized_host = nil
+      @normalized_port = NONE
+      @path = EMPTY_STR
+      @normalized_path = nil
+      @normalized_query = NONE
+      @fragment = nil
+      @normalized_fragment = NONE
+      @query = nil
+    end
+
+    NONE = Module.new.freeze
+
+    private_constant :NONE
   end
 end