addressable 2.8.1 → 2.8.5

data/lib/addressable/uri.rb

@@ -53,7 +53,7 @@ module Addressable
       PCHAR = (UNRESERVED + SUB_DELIMS + "\\:\\@").freeze
       SCHEME = (ALPHA + DIGIT + "\\-\\+\\.").freeze
       HOST = (UNRESERVED + SUB_DELIMS + "\\[\\:\\]").freeze
-      AUTHORITY = (PCHAR + "\\[\\:\\]").freeze
+      AUTHORITY = (PCHAR + "\\[\\]").freeze
       PATH = (PCHAR + "\\/").freeze
       QUERY = (PCHAR + "\\/\\?").freeze
       FRAGMENT = (PCHAR + "\\/\\?").freeze
@@ -117,7 +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)
@@ -138,15 +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(
@@ -189,7 +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
@@ -281,7 +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
@@ -329,13 +329,13 @@ 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
@@ -344,17 +344,13 @@ module Addressable
     ##
     # 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_ENCODING_TABLE = (0..255).map do |byte|
+      format("%02x", byte).freeze
+    end.freeze
 
-    SEQUENCE_UPCASED_PERCENT_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 = (0..255).map do |byte|
+      format("%%%02X", byte).freeze
+    end.freeze
 
     ##
     # Percent encodes a URI component.
@@ -421,16 +417,17 @@ 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_UPCASED_PERCENT_ENCODING_TABLE[sequence]
+      component = component.gsub(character_class) do |char|
+        SEQUENCE_UPCASED_PERCENT_ENCODING_TABLE[char.ord]
       end
       if upcase_encoded.length > 0
-        upcase_encoded_chars = upcase_encoded.chars.map do |char|
-          SEQUENCE_ENCODING_TABLE[char]
+        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
 
@@ -481,7 +478,7 @@ module Addressable
         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
@@ -560,10 +557,9 @@ module Addressable
         leave_re = if leave_encoded.length > 0
           character_class = "#{character_class}%" unless character_class.include?('%')
 
-          "|%(?!#{leave_encoded.chars.flat_map do |char|
-            seq = SEQUENCE_ENCODING_TABLE[char]
-            [seq.upcase, seq.downcase]
-          end.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 = if leave_re
@@ -579,7 +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
         )
@@ -687,8 +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
@@ -836,7 +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]
@@ -851,7 +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
 
     ##
@@ -878,9 +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.
@@ -888,8 +884,8 @@ 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(
@@ -920,7 +916,7 @@ module Addressable
       @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
@@ -931,9 +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.
@@ -941,8 +935,8 @@ 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
@@ -970,14 +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
@@ -988,9 +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.
@@ -998,8 +990,8 @@ 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
@@ -1026,17 +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
@@ -1066,8 +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
@@ -1105,7 +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
@@ -1116,9 +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.
@@ -1161,8 +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
@@ -1293,14 +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
@@ -1348,16 +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
@@ -1384,9 +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.
@@ -1394,8 +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
@@ -1426,8 +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
@@ -1528,9 +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 = /^(?!\/)[^\/:]*:.*$/
     ##
@@ -1579,7 +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
@@ -1609,9 +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.
@@ -1619,8 +1601,8 @@ 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!("\\;", "")
@@ -1652,7 +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
 
@@ -1814,9 +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.
@@ -1824,8 +1804,8 @@ 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::NormalizeCharacterClasses::FRAGMENT
@@ -1848,7 +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
@@ -2014,7 +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
@@ -2408,7 +2388,27 @@ module Addressable
       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
@@ -2507,11 +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
@@ -2543,8 +2539,8 @@ 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
 
     ##
@@ -2556,5 +2552,40 @@ module Addressable
         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

data/lib/addressable/version.rb

@@ -23,7 +23,7 @@ if !defined?(Addressable::VERSION)
     module VERSION
       MAJOR = 2
       MINOR = 8
-      TINY  = 1
+      TINY  = 5
 
       STRING = [MAJOR, MINOR, TINY].join('.')
     end

data/spec/addressable/idna_spec.rb

@@ -38,6 +38,12 @@ shared_examples_for "converting from unicode to ASCII" do
     )).to eq("www.xn--8ws00zhy3a.com")
   end
 
+  it "also accepts unicode strings encoded as ascii-8bit" do
+    expect(Addressable::IDNA.to_ascii(
+      "www.詹姆斯.com".b
+    )).to eq("www.xn--8ws00zhy3a.com")
+  end
+
   it "should convert 'www.Iñtërnâtiônàlizætiøn.com' correctly" do
     "www.Iñtërnâtiônàlizætiøn.com"
     expect(Addressable::IDNA.to_ascii(
@@ -249,11 +255,6 @@ shared_examples_for "converting from ASCII to unicode" do
       "example..host"
     )).to eq("example..host")
   end
-
-  it "should normalize 'string' correctly" do
-    expect(Addressable::IDNA.unicode_normalize_kc(:'string')).to eq("string")
-    expect(Addressable::IDNA.unicode_normalize_kc("string")).to eq("string")
-  end
 end
 
 describe Addressable::IDNA, "when using the pure-Ruby implementation" do