addressable 2.3.7 → 2.7.0

data/lib/addressable/template.rb

@@ -1,6 +1,8 @@
+# frozen_string_literal: true
+
 # encoding:utf-8
 #--
-# Copyright (C) 2006-2013 Bob Aman
+# Copyright (C) Bob Aman
 #
 #    Licensed under the Apache License, Version 2.0 (the "License");
 #    you may not use this file except in compliance with the License.
@@ -234,7 +236,18 @@ module Addressable
       if !pattern.respond_to?(:to_str)
         raise TypeError, "Can't convert #{pattern.class} into String."
       end
-      @pattern = pattern.to_str.freeze
+      @pattern = pattern.to_str.dup.freeze
+    end
+
+    ##
+    # Freeze URI, initializing instance variables.
+    #
+    # @return [Addressable::URI] The frozen URI object.
+    def freeze
+      self.variables
+      self.variable_defaults
+      self.named_captures
+      super
     end
 
     ##
@@ -477,6 +490,8 @@ module Addressable
     # @param [Hash] mapping The mapping that corresponds to the pattern.
     # @param [#validate, #transform] processor
     #   An optional processor object may be supplied.
+    # @param [Boolean] normalize_values
+    #   Optional flag to enable/disable unicode normalization. Default: true
     #
     # The object should respond to either the <tt>validate</tt> or
     # <tt>transform</tt> messages or both. Both the <tt>validate</tt> and
@@ -507,11 +522,11 @@ module Addressable
     #     "http://example.com/{?one,two,three}/"
     #   ).partial_expand({"one" => "1", "three" => 3}).pattern
     #   #=> "http://example.com/?one=1{&two}&three=3"
-    def partial_expand(mapping, processor=nil)
+    def partial_expand(mapping, processor=nil, normalize_values=true)
       result = self.pattern.dup
       mapping = normalize_keys(mapping)
       result.gsub!( EXPRESSION ) do |capture|
-        transform_partial_capture(mapping, capture, processor)
+        transform_partial_capture(mapping, capture, processor, normalize_values)
       end
       return Addressable::Template.new(result)
     end
@@ -522,6 +537,8 @@ module Addressable
     # @param [Hash] mapping The mapping that corresponds to the pattern.
     # @param [#validate, #transform] processor
     #   An optional processor object may be supplied.
+    # @param [Boolean] normalize_values
+    #   Optional flag to enable/disable unicode normalization. Default: true
     #
     # The object should respond to either the <tt>validate</tt> or
     # <tt>transform</tt> messages or both. Both the <tt>validate</tt> and
@@ -572,11 +589,11 @@ module Addressable
     #     ExampleProcessor
     #   ).to_str
     #   #=> Addressable::Template::InvalidTemplateValueError
-    def expand(mapping, processor=nil)
+    def expand(mapping, processor=nil, normalize_values=true)
       result = self.pattern.dup
       mapping = normalize_keys(mapping)
       result.gsub!( EXPRESSION ) do |capture|
-        transform_capture(mapping, capture, processor)
+        transform_capture(mapping, capture, processor, normalize_values)
       end
       return Addressable::URI.parse(result)
     end
@@ -592,6 +609,7 @@ module Addressable
       @variables ||= ordered_variable_defaults.map { |var, val| var }.uniq
     end
     alias_method :keys, :variables
+    alias_method :names, :variables
 
     ##
     # Returns a mapping of variables to their default values specified
@@ -603,9 +621,75 @@ module Addressable
         Hash[*ordered_variable_defaults.reject { |k, v| v.nil? }.flatten]
     end
 
+    ##
+    # Coerces a template into a `Regexp` object. This regular expression will
+    # behave very similarly to the actual template, and should match the same
+    # URI values, but it cannot fully handle, for example, values that would
+    # extract to an `Array`.
+    #
+    # @return [Regexp] A regular expression which should match the template.
+    def to_regexp
+      _, source = parse_template_pattern(pattern)
+      Regexp.new(source)
+    end
+
+    ##
+    # Returns the source of the coerced `Regexp`.
+    #
+    # @return [String] The source of the `Regexp` given by {#to_regexp}.
+    #
+    # @api private
+    def source
+      self.to_regexp.source
+    end
+
+    ##
+    # Returns the named captures of the coerced `Regexp`.
+    #
+    # @return [Hash] The named captures of the `Regexp` given by {#to_regexp}.
+    #
+    # @api private
+    def named_captures
+      self.to_regexp.named_captures
+    end
+
+    ##
+    # Generates a route result for a given set of parameters.
+    # Should only be used by rack-mount.
+    #
+    # @param params [Hash] The set of parameters used to expand the template.
+    # @param recall [Hash] Default parameters used to expand the template.
+    # @param options [Hash] Either a `:processor` or a `:parameterize` block.
+    #
+    # @api private
+    def generate(params={}, recall={}, options={})
+      merged = recall.merge(params)
+      if options[:processor]
+        processor = options[:processor]
+      elsif options[:parameterize]
+        # TODO: This is sending me into fits trying to shoe-horn this into
+        # the existing API. I think I've got this backwards and processors
+        # should be a set of 4 optional blocks named :validate, :transform,
+        # :match, and :restore. Having to use a singleton here is a huge
+        # code smell.
+        processor = Object.new
+        class <<processor
+          attr_accessor :block
+          def transform(name, value)
+            block.call(name, value)
+          end
+        end
+        processor.block = options[:parameterize]
+      else
+        processor = nil
+      end
+      result = self.expand(merged, processor)
+      result.to_s if result
+    end
+
   private
     def ordered_variable_defaults
-      @ordered_variable_defaults ||= (
+      @ordered_variable_defaults ||= begin
         expansions, _ = parse_template_pattern(pattern)
         expansions.map do |capture|
           _, _, varlist = *capture.match(EXPRESSION)
@@ -613,7 +697,7 @@ module Addressable
             varspec[VARSPEC, 1]
           end
         end.flatten
-      )
+      end
     end
 
 
@@ -626,6 +710,8 @@ module Addressable
     #   The expression to expand
     # @param [#validate, #transform] processor
     #   An optional processor object may be supplied.
+    # @param [Boolean] normalize_values
+    #   Optional flag to enable/disable unicode normalization. Default: true
     #
     # The object should respond to either the <tt>validate</tt> or
     # <tt>transform</tt> messages or both. Both the <tt>validate</tt> and
@@ -640,21 +726,36 @@ module Addressable
     # after sending the value to the transform method.
     #
     # @return [String] The expanded expression
-    def transform_partial_capture(mapping, capture, processor = nil)
+    def transform_partial_capture(mapping, capture, processor = nil,
+                                  normalize_values = true)
       _, operator, varlist = *capture.match(EXPRESSION)
-      is_first = true
-      varlist.split(',').inject('') do |acc, varspec|
-        _, name, _ = *varspec.match(VARSPEC)
-        value = mapping[name]
-        if value
-          operator = '&' if !is_first && operator == '?'
-          acc << transform_capture(mapping, "{#{operator}#{varspec}}", processor)
-        else
-          operator = '&' if !is_first && operator == '?'
-          acc << "{#{operator}#{varspec}}"
-        end
-        is_first = false
-        acc
+
+      vars = varlist.split(",")
+
+      if operator == "?"
+        # partial expansion of form style query variables sometimes requires a
+        # slight reordering of the variables to produce a valid url.
+        first_to_expand = vars.find { |varspec|
+          _, name, _ =  *varspec.match(VARSPEC)
+          mapping.key?(name) && !mapping[name].nil?
+        }
+
+        vars = [first_to_expand] + vars.reject {|varspec| varspec == first_to_expand}  if first_to_expand
+      end
+
+      vars.
+        inject("".dup) do |acc, varspec|
+          _, name, _ =  *varspec.match(VARSPEC)
+          next_val = if mapping.key? name
+                       transform_capture(mapping, "{#{operator}#{varspec}}",
+                                         processor, normalize_values)
+                     else
+                       "{#{operator}#{varspec}}"
+                     end
+          # If we've already expanded at least one '?' operator with non-empty
+          # value, change to '&'
+          operator = "&" if (operator == "?") && (next_val != "")
+          acc << next_val
       end
     end
 
@@ -667,6 +768,9 @@ module Addressable
     #   The expression to replace
     # @param [#validate, #transform] processor
     #   An optional processor object may be supplied.
+    # @param [Boolean] normalize_values
+    #   Optional flag to enable/disable unicode normalization. Default: true
+    #
     #
     # The object should respond to either the <tt>validate</tt> or
     # <tt>transform</tt> messages or both. Both the <tt>validate</tt> and
@@ -681,7 +785,8 @@ module Addressable
     # after sending the value to the transform method.
     #
     # @return [String] The expanded expression
-    def transform_capture(mapping, capture, processor=nil)
+    def transform_capture(mapping, capture, processor=nil,
+                          normalize_values=true)
       _, operator, varlist = *capture.match(EXPRESSION)
       return_value = varlist.split(',').inject([]) do |acc, varspec|
         _, name, modifier = *varspec.match(VARSPEC)
@@ -701,7 +806,7 @@ module Addressable
               "Can't convert #{value.class} into String or Array."
           end
 
-          value = normalize_value(value)
+          value = normalize_value(value) if normalize_values
 
           if processor == nil || !processor.respond_to?(:transform)
             # Handle percent escaping
@@ -764,7 +869,9 @@ module Addressable
             end
             if processor.respond_to?(:transform)
               transformed_value = processor.transform(name, value)
-              transformed_value = normalize_value(transformed_value)
+              if normalize_values
+                transformed_value = normalize_value(transformed_value)
+              end
             end
           end
           acc << [name, transformed_value]
@@ -899,7 +1006,7 @@ module Addressable
 
           result = processor && processor.respond_to?(:match) ? processor.match(name) : nil
           if result
-            "(#{ result })"
+            "(?<#{name}>#{ result })"
           else
             group = case operator
             when '+'
@@ -920,9 +1027,9 @@ module Addressable
               "#{ UNRESERVED }*?"
             end
             if modifier == '*'
-              "(#{group}(?:#{joiner}?#{group})*)?"
+              "(?<#{name}>#{group}(?:#{joiner}?#{group})*)?"
             else
-              "(#{group})?"
+              "(?<#{name}>#{group})?"
             end
           end
         end.join("#{joiner}?")

data/lib/addressable/uri.rb

@@ -1,6 +1,8 @@
+# frozen_string_literal: true
+
 # encoding:utf-8
 #--
-# Copyright (C) 2006-2013 Bob Aman
+# Copyright (C) Bob Aman
 #
 #    Licensed under the Apache License, Version 2.0 (the "License");
 #    you may not use this file except in compliance with the License.
@@ -18,6 +20,7 @@
 
 require "addressable/version"
 require "addressable/idna"
+require "public_suffix"
 
 ##
 # Addressable is a library for processing links and URIs.
@@ -44,6 +47,7 @@ module Addressable
       UNRESERVED = ALPHA + DIGIT + "\\-\\.\\_\\~"
       PCHAR = UNRESERVED + SUB_DELIMS + "\\:\\@"
       SCHEME = ALPHA + DIGIT + "\\-\\+\\."
+      HOST = UNRESERVED + SUB_DELIMS + "\\[\\:\\]"
       AUTHORITY = PCHAR
       PATH = PCHAR + "\\/"
       QUERY = PCHAR + "\\/\\?"
@@ -120,9 +124,9 @@ module Addressable
           user = userinfo.strip[/^([^:]*):?/, 1]
           password = userinfo.strip[/:(.*)$/, 1]
         end
-        host = authority.gsub(
+        host = authority.sub(
           /^([^\[\]]*)@/, EMPTY_STR
-        ).gsub(
+        ).sub(
           /:([^:@\[\]]*?)$/, EMPTY_STR
         )
         port = authority[/:([^:@\[\]]*?)$/, 1]
@@ -175,33 +179,50 @@ module Addressable
         raise TypeError, "Can't convert #{uri.class} into String."
       end
       # Otherwise, convert to a String
-      uri = uri.to_str.dup
+      uri = uri.to_str.dup.strip
       hints = {
         :scheme => "http"
       }.merge(hints)
       case uri
-      when /^http:\/+/
-        uri.gsub!(/^http:\/+/, "http://")
-      when /^https:\/+/
-        uri.gsub!(/^https:\/+/, "https://")
-      when /^feed:\/+http:\/+/
-        uri.gsub!(/^feed:\/+http:\/+/, "feed:http://")
-      when /^feed:\/+/
-        uri.gsub!(/^feed:\/+/, "feed://")
-      when /^file:\/+/
-        uri.gsub!(/^file:\/+/, "file:///")
+      when /^http:\//i
+        uri.sub!(/^http:\/+/i, "http://")
+      when /^https:\//i
+        uri.sub!(/^https:\/+/i, "https://")
+      when /^feed:\/+http:\//i
+        uri.sub!(/^feed:\/+http:\/+/i, "feed:http://")
+      when /^feed:\//i
+        uri.sub!(/^feed:\/+/i, "feed://")
+      when %r[^file:/{4}]i
+        uri.sub!(%r[^file:/+]i, "file:////")
+      when %r[^file://localhost/]i
+        uri.sub!(%r[^file://localhost/+]i, "file:///")
+      when %r[^file:/+]i
+        uri.sub!(%r[^file:/+]i, "file:///")
       when /^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}/
-        uri.gsub!(/^/, hints[:scheme] + "://")
+        uri.sub!(/^/, hints[:scheme] + "://")
+      when /\A\d+\..*:\d+\z/
+        uri = "#{hints[:scheme]}://#{uri}"
+      end
+      match = uri.match(URIREGEX)
+      fragments = match.captures
+      authority = fragments[3]
+      if authority && authority.length > 0
+        new_authority = authority.tr("\\", "/").gsub(" ", "%20")
+        # NOTE: We want offset 4, not 3!
+        offset = match.offset(4)
+        uri = uri.dup
+        uri[offset[0]...offset[1]] = new_authority
       end
       parsed = self.parse(uri)
       if parsed.scheme =~ /^[^\/?#\.]+\.[^\/?#]+$/
         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.gsub(
+            new_path = parsed.path.sub(
               Regexp.new("^" + Regexp.escape(new_host)), EMPTY_STR)
             parsed.host = new_host
             parsed.path = new_path
@@ -252,24 +273,24 @@ module Addressable
       # Otherwise, convert to a String
       path = path.to_str.strip
 
-      path.gsub!(/^file:\/?\/?/, EMPTY_STR) if path =~ /^file:\/?\/?/
+      path.sub!(/^file:\/?\/?/, EMPTY_STR) if path =~ /^file:\/?\/?/
       path = SLASH + path if path =~ /^([a-zA-Z])[\|:]/
       uri = self.parse(path)
 
       if uri.scheme == nil
         # Adjust windows-style uris
-        uri.path.gsub!(/^\/?([a-zA-Z])[\|:][\\\/]/) do
+        uri.path.sub!(/^\/?([a-zA-Z])[\|:][\\\/]/) do
           "/#{$1.downcase}:/"
         end
-        uri.path.gsub!(/\\/, SLASH)
-        if File.exists?(uri.path) &&
+        uri.path.tr!("\\", SLASH)
+        if File.exist?(uri.path) &&
             File.stat(uri.path).directory?
-          uri.path.gsub!(/\/$/, 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
@@ -306,6 +327,21 @@ module Addressable
       return result
     end
 
+    ##
+    # Tables used to optimize encoding operations in `self.encode_component`
+    # and `self.normalize_component`
+    SEQUENCE_ENCODING_TABLE = Hash.new do |hash, sequence|
+      hash[sequence] = sequence.unpack("C*").map do |c|
+        format("%02x", c)
+      end.join
+    end
+
+    SEQUENCE_UPCASED_PERCENT_ENCODING_TABLE = Hash.new do |hash, sequence|
+      hash[sequence] = sequence.unpack("C*").map do |c|
+        format("%%%02X", c)
+      end.join
+    end
+
     ##
     # Percent encodes a URI component.
     #
@@ -366,20 +402,20 @@ module Addressable
       if character_class.kind_of?(String)
         character_class = /[^#{character_class}]/
       end
-      if component.respond_to?(:force_encoding)
-        # We can't perform regexps on invalid UTF sequences, but
-        # here we need to, so switch to ASCII.
-        component = component.dup
-        component.force_encoding(Encoding::ASCII_8BIT)
-      end
+      # We can't perform regexps on invalid UTF sequences, but
+      # here we need to, so switch to ASCII.
+      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
+        SEQUENCE_UPCASED_PERCENT_ENCODING_TABLE[sequence]
       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.chars.map do |char|
+          SEQUENCE_ENCODING_TABLE[char]
+        end
+        component = component.gsub(/%(#{upcase_encoded_chars.join('|')})/,
+                                   &:upcase)
       end
       return component
     end
@@ -426,14 +462,14 @@ module Addressable
       end
       uri = uri.dup
       # Seriously, only use UTF-8. I'm really not kidding!
-      uri.force_encoding("utf-8") if uri.respond_to?(:force_encoding)
-      leave_encoded.force_encoding("utf-8") if leave_encoded.respond_to?(:force_encoding)
+      uri.force_encoding("utf-8")
+      leave_encoded = leave_encoded.dup.force_encoding("utf-8")
       result = uri.gsub(/%[0-9a-f]{2}/iu) do |sequence|
         c = sequence[1..3].to_i(16).chr
-        c.force_encoding("utf-8") if c.respond_to?(:force_encoding)
+        c.force_encoding("utf-8")
         leave_encoded.include?(c) ? sequence : c
       end
-      result.force_encoding("utf-8") if result.respond_to?(:force_encoding)
+      result.force_encoding("utf-8")
       if return_type == String
         return result
       elsif return_type == ::Addressable::URI
@@ -513,19 +549,17 @@ module Addressable
           character_class = "#{character_class}%" unless character_class.include?('%')
 
           "|%(?!#{leave_encoded.chars.map do |char|
-            seq = char.unpack('C*').map { |c| '%02x' % c }.join
+            seq = SEQUENCE_ENCODING_TABLE[char]
             [seq.upcase, seq.downcase]
           end.flatten.join('|')})"
         end
 
         character_class = /[^#{character_class}]#{leave_re}/
       end
-      if component.respond_to?(:force_encoding)
-        # We can't perform regexps on invalid UTF sequences, but
-        # here we need to, so switch to ASCII.
-        component = component.dup
-        component.force_encoding(Encoding::ASCII_8BIT)
-      end
+      # We can't perform regexps on invalid UTF sequences, but
+      # here we need to, so switch to ASCII.
+      component = component.dup
+      component.force_encoding(Encoding::ASCII_8BIT)
       unencoded = self.unencode_component(component, String, leave_encoded)
       begin
         encoded = self.encode_component(
@@ -536,9 +570,7 @@ module Addressable
       rescue ArgumentError
         encoded = self.encode_component(unencoded)
       end
-      if encoded.respond_to?(:force_encoding)
-        encoded.force_encoding(Encoding::UTF_8)
-      end
+      encoded.force_encoding(Encoding::UTF_8)
       return encoded
     end
 
@@ -720,9 +752,9 @@ module Addressable
           ).gsub("%20", "+")
         ]
       end
-      return (escaped_form_values.map do |(key, value)|
+      return escaped_form_values.map do |(key, value)|
         "#{key}=#{value}"
-      end).join("&")
+      end.join("&")
     end
 
     ##
@@ -831,7 +863,7 @@ module Addressable
     #
     # @return [String] The scheme component.
     def scheme
-      return instance_variable_defined?(:@scheme) ? @scheme : nil
+      return defined?(@scheme) ? @scheme : nil
     end
 
     ##
@@ -839,16 +871,20 @@ module Addressable
     #
     # @return [String] The scheme component, normalized.
     def normalized_scheme
-      self.scheme && @normalized_scheme ||= (begin
+      return nil unless self.scheme
+      @normalized_scheme ||= begin
         if self.scheme =~ /^\s*ssh\+svn\s*$/i
-          "svn+ssh"
+          "svn+ssh".dup
         else
           Addressable::URI.normalize_component(
             self.scheme.strip.downcase,
             Addressable::URI::CharacterClasses::SCHEME
           )
         end
-      end)
+      end
+      # All normalized values should be UTF-8
+      @normalized_scheme.force_encoding(Encoding::UTF_8) if @normalized_scheme
+      @normalized_scheme
     end
 
     ##
@@ -861,16 +897,15 @@ module Addressable
       elsif new_scheme
         new_scheme = new_scheme.to_str
       end
-      if new_scheme && new_scheme !~ /[a-z][a-z0-9\.\+\-]*/i
-        raise InvalidURIError, "Invalid scheme format."
+      if new_scheme && new_scheme !~ /\A[a-z][a-z0-9\.\+\-]*\z/i
+        raise InvalidURIError, "Invalid scheme format: #{new_scheme}"
       end
       @scheme = new_scheme
       @scheme = nil if @scheme.to_s.strip.empty?
 
-      # Reset dependant values
-      @normalized_scheme = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@normalized_scheme) if defined?(@normalized_scheme)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -881,7 +916,7 @@ module Addressable
     #
     # @return [String] The user component.
     def user
-      return instance_variable_defined?(:@user) ? @user : nil
+      return defined?(@user) ? @user : nil
     end
 
     ##
@@ -889,7 +924,9 @@ module Addressable
     #
     # @return [String] The user component, normalized.
     def normalized_user
-      self.user && @normalized_user ||= (begin
+      return nil unless self.user
+      return @normalized_user if defined?(@normalized_user)
+      @normalized_user ||= begin
         if normalized_scheme =~ /https?/ && self.user.strip.empty? &&
             (!self.password || self.password.strip.empty?)
           nil
@@ -899,7 +936,10 @@ module Addressable
             Addressable::URI::CharacterClasses::UNRESERVED
           )
         end
-      end)
+      end
+      # All normalized values should be UTF-8
+      @normalized_user.force_encoding(Encoding::UTF_8) if @normalized_user
+      @normalized_user
     end
 
     ##
@@ -917,13 +957,12 @@ module Addressable
         @user = EMPTY_STR if @user.nil?
       end
 
-      # Reset dependant values
-      @userinfo = nil
-      @normalized_userinfo = nil
-      @authority = nil
-      @normalized_user = nil
-      @uri_string = nil
-      @hash = 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_user) if defined?(@normalized_user)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -934,7 +973,7 @@ module Addressable
     #
     # @return [String] The password component.
     def password
-      return instance_variable_defined?(:@password) ? @password : nil
+      return defined?(@password) ? @password : nil
     end
 
     ##
@@ -942,7 +981,9 @@ module Addressable
     #
     # @return [String] The password component, normalized.
     def normalized_password
-      self.password && @normalized_password ||= (begin
+      return nil unless self.password
+      return @normalized_password if defined?(@normalized_password)
+      @normalized_password ||= begin
         if self.normalized_scheme =~ /https?/ && self.password.strip.empty? &&
             (!self.user || self.user.strip.empty?)
           nil
@@ -952,7 +993,12 @@ module Addressable
             Addressable::URI::CharacterClasses::UNRESERVED
           )
         end
-      end)
+      end
+      # All normalized values should be UTF-8
+      if @normalized_password
+        @normalized_password.force_encoding(Encoding::UTF_8)
+      end
+      @normalized_password
     end
 
     ##
@@ -972,13 +1018,12 @@ module Addressable
         @user = EMPTY_STR if @user.nil?
       end
 
-      # Reset dependant values
-      @userinfo = nil
-      @normalized_userinfo = nil
-      @authority = nil
-      @normalized_password = nil
-      @uri_string = nil
-      @hash = 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_password) if defined?(@normalized_password)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -992,13 +1037,13 @@ module Addressable
     def userinfo
       current_user = self.user
       current_password = self.password
-      (current_user || current_password) && @userinfo ||= (begin
+      (current_user || current_password) && @userinfo ||= begin
         if current_user && current_password
           "#{current_user}:#{current_password}"
         elsif current_user && !current_password
           "#{current_user}"
         end
-      end)
+      end
     end
 
     ##
@@ -1006,17 +1051,24 @@ module Addressable
     #
     # @return [String] The userinfo component, normalized.
     def normalized_userinfo
-      self.userinfo && @normalized_userinfo ||= (begin
+      return nil unless self.userinfo
+      return @normalized_userinfo if defined?(@normalized_userinfo)
+      @normalized_userinfo ||= begin
         current_user = self.normalized_user
         current_password = self.normalized_password
         if !current_user && !current_password
           nil
         elsif current_user && current_password
-          "#{current_user}:#{current_password}"
+          "#{current_user}:#{current_password}".dup
         elsif current_user && !current_password
-          "#{current_user}"
+          "#{current_user}".dup
         end
-      end)
+      end
+      # All normalized values should be UTF-8
+      if @normalized_userinfo
+        @normalized_userinfo.force_encoding(Encoding::UTF_8)
+      end
+      @normalized_userinfo
     end
 
     ##
@@ -1040,10 +1092,9 @@ module Addressable
       self.password = new_password
       self.user = new_user
 
-      # Reset dependant values
-      @authority = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@authority) if defined?(@authority)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -1054,7 +1105,7 @@ module Addressable
     #
     # @return [String] The host component.
     def host
-      return instance_variable_defined?(:@host) ? @host : nil
+      return defined?(@host) ? @host : nil
     end
 
     ##
@@ -1062,7 +1113,8 @@ module Addressable
     #
     # @return [String] The host component, normalized.
     def normalized_host
-      self.host && @normalized_host ||= (begin
+      return nil unless self.host
+      @normalized_host ||= begin
         if !self.host.strip.empty?
           result = ::Addressable::IDNA.to_ascii(
             URI.unencode_component(self.host.strip.downcase)
@@ -1071,11 +1123,17 @@ module Addressable
             # Single trailing dots are unnecessary.
             result = result[0...-1]
           end
+          result = Addressable::URI.normalize_component(
+            result,
+            CharacterClasses::HOST)
           result
         else
-          EMPTY_STR
+          EMPTY_STR.dup
         end
-      end)
+      end
+      # All normalized values should be UTF-8
+      @normalized_host.force_encoding(Encoding::UTF_8) if @normalized_host
+      @normalized_host
     end
 
     ##
@@ -1088,19 +1146,10 @@ module Addressable
       end
       @host = new_host ? new_host.to_str : nil
 
-      unreserved = CharacterClasses::UNRESERVED
-      sub_delims = CharacterClasses::SUB_DELIMS
-      if @host != nil && (@host =~ /[<>{}\/\?\#\@]/ ||
-          (@host[/^\[(.*)\]$/, 1] != nil && @host[/^\[(.*)\]$/, 1] !~
-          Regexp.new("^[#{unreserved}#{sub_delims}:]*$")))
-        raise InvalidURIError, "Invalid character in host: '#{@host.to_s}'"
-      end
-
-      # Reset dependant values
-      @authority = nil
-      @normalized_host = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@authority) if defined?(@authority)
+      remove_instance_variable(:@normalized_host) if defined?(@normalized_host)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -1126,7 +1175,10 @@ module Addressable
     #
     # @param [String, #to_str] new_hostname The new hostname for this URI.
     def hostname=(new_hostname)
-      if new_hostname && !new_hostname.respond_to?(:to_str)
+      if new_hostname &&
+          (new_hostname.respond_to?(:ipv4?) || new_hostname.respond_to?(:ipv6?))
+        new_hostname = new_hostname.to_s
+      elsif new_hostname && !new_hostname.respond_to?(:to_str)
         raise TypeError, "Can't convert #{new_hostname.class} into String."
       end
       v = new_hostname ? new_hostname.to_str : nil
@@ -1134,14 +1186,41 @@ module Addressable
       self.host = v
     end
 
+    ##
+    # Returns the top-level domain for this host.
+    #
+    # @example
+    #   Addressable::URI.parse("http://www.example.co.uk").tld # => "co.uk"
+    def tld
+      PublicSuffix.parse(self.host, ignore_private: true).tld
+    end
+
+    ##
+    # Sets the top-level domain for this URI.
+    #
+    # @param [String, #to_str] new_tld The new top-level domain.
+    def tld=(new_tld)
+      replaced_tld = host.sub(/#{tld}\z/, new_tld)
+      self.host = PublicSuffix::Domain.new(replaced_tld).to_s
+    end
+
+    ##
+    # Returns the public suffix domain for this host.
+    #
+    # @example
+    #   Addressable::URI.parse("http://www.example.co.uk").domain # => "example.co.uk"
+    def domain
+      PublicSuffix.domain(self.host, ignore_private: true)
+    end
+
     ##
     # The authority component for this URI.
     # Combines the user, password, host, and port components.
     #
     # @return [String] The authority component.
     def authority
-      self.host && @authority ||= (begin
-        authority = ""
+      self.host && @authority ||= begin
+        authority = String.new
         if self.userinfo != nil
           authority << "#{self.userinfo}@"
         end
@@ -1150,7 +1229,7 @@ module Addressable
           authority << ":#{self.port}"
         end
         authority
-      end)
+      end
     end
 
     ##
@@ -1158,8 +1237,9 @@ module Addressable
     #
     # @return [String] The authority component, normalized.
     def normalized_authority
-      self.authority && @normalized_authority ||= (begin
-        authority = ""
+      return nil unless self.authority
+      @normalized_authority ||= begin
+        authority = String.new
         if self.normalized_userinfo != nil
           authority << "#{self.normalized_userinfo}@"
         end
@@ -1168,7 +1248,12 @@ module Addressable
           authority << ":#{self.normalized_port}"
         end
         authority
-      end)
+      end
+      # All normalized values should be UTF-8
+      if @normalized_authority
+        @normalized_authority.force_encoding(Encoding::UTF_8)
+      end
+      @normalized_authority
     end
 
     ##
@@ -1186,9 +1271,9 @@ module Addressable
           new_user = new_userinfo.strip[/^([^:]*):?/, 1]
           new_password = new_userinfo.strip[/:(.*)$/, 1]
         end
-        new_host = new_authority.gsub(
+        new_host = new_authority.sub(
           /^([^\[\]]*)@/, EMPTY_STR
-        ).gsub(
+        ).sub(
           /:([^:@\[\]]*?)$/, EMPTY_STR
         )
         new_port =
@@ -1201,11 +1286,10 @@ module Addressable
       self.host = defined?(new_host) ? new_host : nil
       self.port = defined?(new_port) ? new_port : nil
 
-      # Reset dependant values
-      @userinfo = nil
-      @normalized_userinfo = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@userinfo) if defined?(@userinfo)
+      remove_instance_variable(:@normalized_userinfo) if defined?(@normalized_userinfo)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -1217,18 +1301,55 @@ module Addressable
     #
     # @return [String] The serialized origin.
     def origin
-      return (if self.scheme && self.authority
+      if self.scheme && self.authority
         if self.normalized_port
-          (
-            "#{self.normalized_scheme}://#{self.normalized_host}" +
-            ":#{self.normalized_port}"
-          )
+          "#{self.normalized_scheme}://#{self.normalized_host}" +
+          ":#{self.normalized_port}"
         else
           "#{self.normalized_scheme}://#{self.normalized_host}"
         end
       else
         "null"
-      end)
+      end
+    end
+
+    ##
+    # Sets the origin for this URI, serialized to ASCII, as per
+    # RFC 6454, section 6.2. This assignment will reset the `userinfo`
+    # component.
+    #
+    # @param [String, #to_str] new_origin The new origin component.
+    def origin=(new_origin)
+      if new_origin
+        if !new_origin.respond_to?(:to_str)
+          raise TypeError, "Can't convert #{new_origin.class} into String."
+        end
+        new_origin = new_origin.to_str
+        new_scheme = new_origin[/^([^:\/?#]+):\/\//, 1]
+        unless new_scheme
+          raise InvalidURIError, 'An origin cannot omit the scheme.'
+        end
+        new_host = new_origin[/:\/\/([^\/?#:]+)/, 1]
+        unless new_host
+          raise InvalidURIError, 'An origin cannot omit the host.'
+        end
+        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.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)
+      remove_composite_values
+
+      # Ensure we haven't created an invalid URI
+      validate()
     end
 
     # Returns an array of known ip-based schemes. These schemes typically
@@ -1252,7 +1373,7 @@ module Addressable
     #
     # @return [Integer] The port component.
     def port
-      return instance_variable_defined?(:@port) ? @port : nil
+      return defined?(@port) ? @port : nil
     end
 
     ##
@@ -1260,10 +1381,14 @@ module Addressable
     #
     # @return [Integer] The port component, normalized.
     def normalized_port
-      if URI.port_mapping[self.normalized_scheme] == self.port
-        nil
-      else
-        self.port
+      return nil unless self.port
+      return @normalized_port if defined?(@normalized_port)
+      @normalized_port ||= begin
+        if URI.port_mapping[self.normalized_scheme] == self.port
+          nil
+        else
+          self.port
+        end
       end
     end
 
@@ -1275,6 +1400,11 @@ module Addressable
       if new_port != nil && new_port.respond_to?(:to_str)
         new_port = Addressable::URI.unencode_component(new_port.to_str)
       end
+
+      if new_port.respond_to?(:valid_encoding?) && !new_port.valid_encoding?
+        raise InvalidURIError, "Invalid encoding in port"
+      end
+
       if new_port != nil && !(new_port.to_s =~ /^\d+$/)
         raise InvalidURIError,
           "Invalid port number: #{new_port.inspect}"
@@ -1283,11 +1413,10 @@ module Addressable
       @port = new_port.to_s.to_i
       @port = nil if @port == 0
 
-      # Reset dependant values
-      @authority = nil
-      @normalized_port = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@authority) if defined?(@authority)
+      remove_instance_variable(:@normalized_port) if defined?(@normalized_port)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -1327,12 +1456,12 @@ module Addressable
     #
     # @return [String] The components that identify a site.
     def site
-      (self.scheme || self.authority) && @site ||= (begin
-        site_string = ""
+      (self.scheme || self.authority) && @site ||= begin
+        site_string = "".dup
         site_string << "#{self.scheme}:" if self.scheme != nil
         site_string << "//#{self.authority}" if self.authority != nil
         site_string
-      end)
+      end
     end
 
     ##
@@ -1345,8 +1474,9 @@ module Addressable
     #
     # @return [String] The normalized components that identify a site.
     def normalized_site
-      self.site && @normalized_site ||= (begin
-        site_string = ""
+      return nil unless self.site
+      @normalized_site ||= begin
+        site_string = "".dup
         if self.normalized_scheme != nil
           site_string << "#{self.normalized_scheme}:"
         end
@@ -1354,7 +1484,10 @@ module Addressable
           site_string << "//#{self.normalized_authority}"
         end
         site_string
-      end)
+      end
+      # All normalized values should be UTF-8
+      @normalized_site.force_encoding(Encoding::UTF_8) if @normalized_site
+      @normalized_site
     end
 
     ##
@@ -1384,7 +1517,7 @@ module Addressable
     #
     # @return [String] The path component.
     def path
-      return instance_variable_defined?(:@path) ? @path : EMPTY_STR
+      return defined?(@path) ? @path : EMPTY_STR
     end
 
     NORMPATH = /^(?!\/)[^\/:]*:.*$/
@@ -1393,7 +1526,7 @@ module Addressable
     #
     # @return [String] The path component, normalized.
     def normalized_path
-      @normalized_path ||= (begin
+      @normalized_path ||= begin
         path = self.path.to_s
         if self.scheme == nil && path =~ NORMPATH
           # Relative paths with colons in the first segment are ambiguous.
@@ -1401,20 +1534,23 @@ module Addressable
         end
         # String#split(delimeter, -1) uses the more strict splitting behavior
         # found by default in Python.
-        result = (path.strip.split(SLASH, -1).map do |segment|
+        result = path.strip.split(SLASH, -1).map do |segment|
           Addressable::URI.normalize_component(
             segment,
             Addressable::URI::CharacterClasses::PCHAR
           )
-        end).join(SLASH)
+        end.join(SLASH)
 
         result = URI.normalize_path(result)
         if result.empty? &&
             ["http", "https", "ftp", "tftp"].include?(self.normalized_scheme)
-          result = SLASH
+          result = SLASH.dup
         end
         result
-      end)
+      end
+      # All normalized values should be UTF-8
+      @normalized_path.force_encoding(Encoding::UTF_8) if @normalized_path
+      @normalized_path
     end
 
     ##
@@ -1430,10 +1566,12 @@ module Addressable
         @path = "/#{@path}"
       end
 
-      # Reset dependant values
-      @normalized_path = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@normalized_path) if defined?(@normalized_path)
+      remove_composite_values
+
+      # Ensure we haven't created an invalid URI
+      validate()
     end
 
     ##
@@ -1442,7 +1580,7 @@ module Addressable
     # @return [String] The path's basename.
     def basename
       # Path cannot be nil
-      return File.basename(self.path).gsub(/;[^\/]*$/, EMPTY_STR)
+      return File.basename(self.path).sub(/;[^\/]*$/, EMPTY_STR)
     end
 
     ##
@@ -1460,7 +1598,7 @@ module Addressable
     #
     # @return [String] The query component.
     def query
-      return instance_variable_defined?(:@query) ? @query : nil
+      return defined?(@query) ? @query : nil
     end
 
     ##
@@ -1468,15 +1606,23 @@ module Addressable
     #
     # @return [String] The query component, normalized.
     def normalized_query(*flags)
-      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.sort! if flags.include?(:sorted)
-      component = (pairs.map do |pair|
-        Addressable::URI.normalize_component(pair, modified_query_class, "+")
-      end).join("&")
-      component == "" ? nil : component
+      return nil unless self.query
+      return @normalized_query if defined?(@normalized_query)
+      @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.delete_if(&:empty?) if flags.include?(:compacted)
+        pairs.sort! if flags.include?(:sorted)
+        component = pairs.map do |pair|
+          Addressable::URI.normalize_component(pair, modified_query_class, "+")
+        end.join("&")
+        component == "" ? nil : component
+      end
+      # All normalized values should be UTF-8
+      @normalized_query.force_encoding(Encoding::UTF_8) if @normalized_query
+      @normalized_query
     end
 
     ##
@@ -1489,10 +1635,9 @@ module Addressable
       end
       @query = new_query ? new_query.to_str : nil
 
-      # Reset dependant values
-      @normalized_query = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@normalized_query) if defined?(@normalized_query)
+      remove_composite_values
     end
 
     ##
@@ -1501,7 +1646,8 @@ module Addressable
     # @param [Class] return_type The return type desired. Value must be either
     #   `Hash` or `Array`.
     #
-    # @return [Hash, Array] The query string parsed as a Hash or Array object.
+    # @return [Hash, Array, nil] The query string parsed as a Hash or Array
+    #   or nil if the query string is blank.
     #
     # @example
     #   Addressable::URI.parse("?one=1&two=2&three=3").query_values
@@ -1510,15 +1656,19 @@ module Addressable
     #   #=> [["one", "two"], ["one", "three"]]
     #   Addressable::URI.parse("?one=two&one=three").query_values(Hash)
     #   #=> {"one" => "three"}
+    #   Addressable::URI.parse("?").query_values
+    #   #=> {}
+    #   Addressable::URI.parse("").query_values
+    #   #=> nil
     def query_values(return_type=Hash)
       empty_accumulator = Array == return_type ? [] : {}
       if return_type != Hash && return_type != Array
         raise ArgumentError, "Invalid return type. Must be Hash or Array."
       end
       return nil if self.query == nil
-      split_query = (self.query.split("&").map do |pair|
+      split_query = self.query.split("&").map do |pair|
         pair.split("=", 2) if pair && !pair.empty?
-      end).compact
+      end.compact
       return split_query.inject(empty_accumulator.dup) do |accu, pair|
         # I'd rather use key/value identifiers instead of array lookups,
         # but in this case I really want to maintain the exact pair structure,
@@ -1529,7 +1679,7 @@ module Addressable
           # 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(/\+/, " "))
+          pair[1] = URI.unencode_component(pair[1].to_str.tr("+", " "))
         end
         if return_type == Hash
           accu[pair[0]] = pair[1]
@@ -1581,7 +1731,7 @@ module Addressable
       end
 
       # new_query_values have form [['key1', 'value1'], ['key2', 'value2']]
-      buffer = ""
+      buffer = "".dup
       new_query_values.each do |key, value|
         encoded_key = URI.encode_component(
           key, CharacterClasses::UNRESERVED
@@ -1611,7 +1761,7 @@ module Addressable
     #
     # @return [String] The request URI required for an HTTP request.
     def request_uri
-      return nil if self.absolute? && self.scheme !~ /^https?$/
+      return nil if self.absolute? && self.scheme !~ /^https?$/i
       return (
         (!self.path.empty? ? self.path : SLASH) +
         (self.query ? "?#{self.query}" : EMPTY_STR)
@@ -1626,21 +1776,20 @@ module Addressable
       if !new_request_uri.respond_to?(:to_str)
         raise TypeError, "Can't convert #{new_request_uri.class} into String."
       end
-      if self.absolute? && self.scheme !~ /^https?$/
+      if self.absolute? && self.scheme !~ /^https?$/i
         raise InvalidURIError,
           "Cannot set an HTTP request URI for a non-HTTP URI."
       end
       new_request_uri = new_request_uri.to_str
-      path_component = new_request_uri[/^([^\?]*)\?(?:.*)$/, 1]
+      path_component = new_request_uri[/^([^\?]*)\??(?:.*)$/, 1]
       query_component = new_request_uri[/^(?:[^\?]*)\?(.*)$/, 1]
       path_component = path_component.to_s
       path_component = (!path_component.empty? ? path_component : SLASH)
       self.path = path_component
       self.query = query_component
 
-      # Reset dependant values
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_composite_values
     end
 
     ##
@@ -1648,7 +1797,7 @@ module Addressable
     #
     # @return [String] The fragment component.
     def fragment
-      return instance_variable_defined?(:@fragment) ? @fragment : nil
+      return defined?(@fragment) ? @fragment : nil
     end
 
     ##
@@ -1656,13 +1805,20 @@ module Addressable
     #
     # @return [String] The fragment component, normalized.
     def normalized_fragment
-      self.fragment && @normalized_fragment ||= (begin
+      return nil unless self.fragment
+      return @normalized_fragment if defined?(@normalized_fragment)
+      @normalized_fragment ||= begin
         component = Addressable::URI.normalize_component(
           self.fragment,
           Addressable::URI::CharacterClasses::FRAGMENT
         )
         component == "" ? nil : component
-      end)
+      end
+      # All normalized values should be UTF-8
+      if @normalized_fragment
+        @normalized_fragment.force_encoding(Encoding::UTF_8)
+      end
+      @normalized_fragment
     end
 
     ##
@@ -1675,10 +1831,9 @@ module Addressable
       end
       @fragment = new_fragment ? new_fragment.to_str : nil
 
-      # Reset dependant values
-      @normalized_fragment = nil
-      @uri_string = nil
-      @hash = nil
+      # Reset dependent values
+      remove_instance_variable(:@normalized_fragment) if defined?(@normalized_fragment)
+      remove_composite_values
 
       # Ensure we haven't created an invalid URI
       validate()
@@ -1781,8 +1936,8 @@ module Addressable
               # Section 5.2.3 of RFC 3986
               #
               # Removes the right-most path segment from the base path.
-              if base_path =~ /\//
-                base_path.gsub!(/\/[^\/]+$/, SLASH)
+              if base_path.include?(SLASH)
+                base_path.sub!(/\/[^\/]+$/, SLASH)
               else
                 base_path = EMPTY_STR
               end
@@ -2099,7 +2254,7 @@ module Addressable
     #
     # @return [Integer] A hash of the URI.
     def hash
-      return @hash ||= (self.to_s.hash * -1)
+      @hash ||= self.to_s.hash * -1
     end
 
     ##
@@ -2182,18 +2337,16 @@ module Addressable
         raise InvalidURIError,
           "Cannot assemble URI string with ambiguous path: '#{self.path}'"
       end
-      @uri_string ||= (begin
-        uri_string = ""
+      @uri_string ||= begin
+        uri_string = String.new
         uri_string << "#{self.scheme}:" if self.scheme != nil
         uri_string << "//#{self.authority}" if self.authority != nil
         uri_string << self.path.to_s
         uri_string << "?#{self.query}" if self.query != nil
         uri_string << "##{self.fragment}" if self.fragment != nil
-        if uri_string.respond_to?(:force_encoding)
-          uri_string.force_encoding(Encoding::UTF_8)
-        end
+        uri_string.force_encoding(Encoding::UTF_8)
         uri_string
-      end)
+      end
     end
 
     ##
@@ -2233,16 +2386,16 @@ 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
     end
 
-  private
+  protected
     SELF_REF = '.'
     PARENT = '..'
 
@@ -2308,6 +2461,19 @@ module Addressable
         raise InvalidURIError,
           "Cannot have a relative path with an authority set: '#{self.to_s}'"
       end
+      if self.path != nil && !self.path.empty? &&
+          self.path[0..1] == SLASH + SLASH && self.authority == nil
+        raise InvalidURIError,
+          "Cannot have a path with two leading slashes " +
+          "without an authority set: '#{self.to_s}'"
+      end
+      unreserved = CharacterClasses::UNRESERVED
+      sub_delims = CharacterClasses::SUB_DELIMS
+      if !self.host.nil? && (self.host =~ /[<>{}\/\\\?\#\@"[[:space:]]]/ ||
+          (self.host[/^\[(.*)\]$/, 1] != nil && self.host[/^\[(.*)\]$/, 1] !~
+          Regexp.new("^[#{unreserved}#{sub_delims}:]*$")))
+        raise InvalidURIError, "Invalid character in host: '#{self.host.to_s}'"
+      end
       return nil
     end
 
@@ -2319,9 +2485,11 @@ module Addressable
     #
     # @return [Addressable::URI] <code>self</code>.
     def replace_self(uri)
-      # Reset dependant values
+      # Reset dependent values
       instance_variables.each do |var|
-        instance_variable_set(var, nil)
+        if instance_variable_defined?(var) && var != :@validation_deferred
+          remove_instance_variable(var)
+        end
       end
 
       @scheme = uri.scheme
@@ -2336,7 +2504,7 @@ module Addressable
     end
 
     ##
-    # Splits path string with "/"(slash).
+    # Splits path string with "/" (slash).
     # It is considered that there is empty string after last slash when
     # path ends with slash.
     #
@@ -2348,5 +2516,14 @@ module Addressable
       splitted << EMPTY_STR if path.end_with? SLASH
       splitted
     end
+
+    ##
+    # Resets composite values for the entire URI
+    #
+    # @api private
+    def remove_composite_values
+      remove_instance_variable(:@uri_string) if defined?(@uri_string)
+      remove_instance_variable(:@hash) if defined?(@hash)
+    end
   end
 end