mail 2.6.6 → 2.8.1

data/lib/mail/utilities.rb

@@ -1,16 +1,15 @@
 # encoding: utf-8
 # frozen_string_literal: true
+require 'mail/constants'
+require 'socket'
+
 module Mail
   module Utilities
-
-    LF   = "\n"
-    CRLF = "\r\n"
-
-    include Constants
+    extend self
 
     # Returns true if the string supplied is free from characters not allowed as an ATOM
     def atom_safe?( str )
-      not ATOM_UNSAFE === str
+      not Constants::ATOM_UNSAFE === str
     end
 
     # If the string supplied has ATOM unsafe characters in it, will return the string quoted
@@ -22,28 +21,38 @@ module Mail
     # If the string supplied has PHRASE unsafe characters in it, will return the string quoted
     # in double quotes, otherwise returns the string unmodified
     def quote_phrase( str )
-      if RUBY_VERSION >= '1.9'
+      if str.respond_to?(:force_encoding)
         original_encoding = str.encoding
-        ascii_str = str.dup.force_encoding('ASCII-8BIT')
-        if (PHRASE_UNSAFE === ascii_str)
+        ascii_str = str.to_s.dup.force_encoding('ASCII-8BIT')
+        if Constants::PHRASE_UNSAFE === ascii_str
           dquote(ascii_str).force_encoding(original_encoding)
         else
           str
         end
       else
-        (PHRASE_UNSAFE === str) ? dquote(str) : str
+        Constants::PHRASE_UNSAFE === str ? dquote(str) : str
       end
     end
 
     # Returns true if the string supplied is free from characters not allowed as a TOKEN
     def token_safe?( str )
-      not TOKEN_UNSAFE === str
+      not Constants::TOKEN_UNSAFE === str
     end
 
     # If the string supplied has TOKEN unsafe characters in it, will return the string quoted
     # in double quotes, otherwise returns the string unmodified
     def quote_token( str )
-      token_safe?( str ) ? str : dquote(str)
+      if str.respond_to?(:force_encoding)
+        original_encoding = str.encoding
+        ascii_str = str.to_s.dup.force_encoding('ASCII-8BIT')
+        if token_safe?( ascii_str )
+          str
+        else
+          dquote(ascii_str).force_encoding(original_encoding)
+        end
+      else
+        token_safe?( str ) ? str : dquote(str)
+      end
     end
 
     # Wraps supplied string in double quotes and applies \-escaping as necessary,
@@ -90,7 +99,6 @@ module Mail
     def unescape( str )
       str.gsub(/\\(.)/, '\1')
     end
-    module_function :unescape
 
     # Wraps a string in parenthesis and escapes any that are in the string itself.
     #
@@ -98,7 +106,7 @@ module Mail
     #
     #  paren( 'This is a string' ) #=> '(This is a string)'
     def paren( str )
-      RubyVer.paren( str )
+      Utilities.paren( str )
     end
 
     # Unwraps a string from being wrapped in parenthesis
@@ -108,8 +116,11 @@ module Mail
     #  str = '(This is a string)'
     #  unparen( str ) #=> 'This is a string'
     def unparen( str )
-      match = str.match(/^\((.*?)\)$/)
-      match ? match[1] : str
+      if str.start_with?('(') && str.end_with?(')')
+        str.slice(1..-2)
+      else
+        str
+      end
     end
 
     # Wraps a string in angle brackets and escapes any that are in the string itself
@@ -118,7 +129,7 @@ module Mail
     #
     #  bracket( 'This is a string' ) #=> '<This is a string>'
     def bracket( str )
-      RubyVer.bracket( str )
+      Utilities.bracket( str )
     end
 
     # Unwraps a string from being wrapped in parenthesis
@@ -128,8 +139,11 @@ module Mail
     #  str = '<This is a string>'
     #  unbracket( str ) #=> 'This is a string'
     def unbracket( str )
-      match = str.match(/^\<(.*?)\>$/)
-      match ? match[1] : str
+      if str.start_with?('<') && str.end_with?('>')
+        str.slice(1..-2)
+      else
+        str
+      end
     end
 
     # Escape parenthesies in a string
@@ -139,7 +153,7 @@ module Mail
     #  str = 'This is (a) string'
     #  escape_paren( str ) #=> 'This is \(a\) string'
     def escape_paren( str )
-      RubyVer.escape_paren( str )
+      Utilities.escape_paren( str )
     end
 
     def uri_escape( str )
@@ -151,7 +165,7 @@ module Mail
     end
 
     def uri_parser
-      @uri_parser ||= URI.const_defined?(:Parser) ? URI::Parser.new : URI
+      @uri_parser ||= URI.const_defined?(:DEFAULT_PARSER) ? URI::DEFAULT_PARSER : URI
     end
 
     # Matches two objects with their to_s values case insensitively
@@ -192,9 +206,9 @@ module Mail
     # Example:
     #
     #  string = :resent_from_field
-    #  dasherize ( string ) #=> 'resent_from_field'
+    #  dasherize( string ) #=> 'resent-from-field'
     def dasherize( str )
-      str.to_s.tr(UNDERSCORE, HYPHEN)
+      str.to_s.tr(Constants::UNDERSCORE, Constants::HYPHEN)
     end
 
     # Swaps out all hyphens (-) for underscores (_) good for stringing to symbols
@@ -205,89 +219,344 @@ module Mail
     #  string = :resent_from_field
     #  underscoreize ( string ) #=> 'resent_from_field'
     def underscoreize( str )
-      str.to_s.downcase.tr(HYPHEN, UNDERSCORE)
+      str.to_s.downcase.tr(Constants::HYPHEN, Constants::UNDERSCORE)
+    end
+
+    def map_lines( str, &block )
+      str.each_line.map(&block)
+    end
+
+    def map_with_index( enum, &block )
+      enum.each_with_index.map(&block)
+    end
+
+    def self.binary_unsafe_to_lf(string) #:nodoc:
+      string.gsub(/\r\n|\r/, Constants::LF)
+    end
+
+    TO_CRLF_REGEX =
+      # This 1.9 only regex can save a reasonable amount of time (~20%)
+      # by not matching "\r\n" so the string is returned unchanged in
+      # the common case.
+      Regexp.new("(?<!\r)\n|\r(?!\n)")
+
+    def self.binary_unsafe_to_crlf(string) #:nodoc:
+      string.gsub(TO_CRLF_REGEX, Constants::CRLF)
+    end
+
+    def self.safe_for_line_ending_conversion?(string) #:nodoc:
+      if string.encoding == Encoding::BINARY
+        string.ascii_only?
+      else
+        string.valid_encoding?
+      end
+    end
+
+    # Convert line endings to \n unless the string is binary. Used for
+    # sendmail delivery and for decoding 8bit Content-Transfer-Encoding.
+    def self.to_lf(string)
+      string = string.to_s
+      if safe_for_line_ending_conversion? string
+        binary_unsafe_to_lf string
+      else
+        string
+      end
     end
 
-    if RUBY_VERSION <= '1.8.6'
+    # Convert line endings to \r\n unless the string is binary. Used for
+    # encoding 8bit and base64 Content-Transfer-Encoding and for convenience
+    # when parsing emails with \n line endings instead of the required \r\n.
+    def self.to_crlf(string)
+      string = string.to_s
+      if safe_for_line_ending_conversion? string
+        binary_unsafe_to_crlf string
+      else
+        string
+      end
+    end
 
-      def map_lines( str, &block )
-        results = []
-        str.each_line do |line|
-          results << yield(line)
+    # Returns true if the object is considered blank.
+    # A blank includes things like '', '   ', nil,
+    # and arrays and hashes that have nothing in them.
+    #
+    # This logic is mostly shared with ActiveSupport's blank?
+    def blank?(value)
+      if value.kind_of?(NilClass)
+        true
+      elsif value.kind_of?(String)
+        value !~ /\S/
+      else
+        value.respond_to?(:empty?) ? value.empty? : !value
+      end
+    end
+
+    def generate_message_id
+      "<#{Mail.random_tag}@#{::Socket.gethostname}.mail>"
+    end
+
+    class StrictCharsetEncoder
+      def encode(string, charset)
+        case charset
+        when /utf-?7/i
+          Mail::Utilities.decode_utf7(string)
+        else
+          string.force_encoding(Mail::Utilities.pick_encoding(charset))
         end
-        results
       end
+    end
 
-      def map_with_index( enum, &block )
-        results = []
-        enum.each_with_index do |token, i|
-          results[i] = yield(token, i)
+    class BestEffortCharsetEncoder
+      def encode(string, charset)
+        case charset
+        when /utf-?7/i
+          Mail::Utilities.decode_utf7(string)
+        else
+          string.force_encoding(pick_encoding(charset))
         end
-        results
       end
 
-    else
+      private
 
-      def map_lines( str, &block )
-        str.each_line.map(&block)
+      def pick_encoding(charset)
+        charset = case charset
+        when /ansi_x3.110-1983/
+          'ISO-8859-1'
+        when /Windows-?1258/i # Windows-1258 is similar to 1252
+          "Windows-1252"
+        else
+          charset
+        end
+        Mail::Utilities.pick_encoding(charset)
       end
+    end
 
-      def map_with_index( enum, &block )
-        enum.each_with_index.map(&block)
-      end
+    class << self
+      attr_accessor :charset_encoder
+    end
+    self.charset_encoder = BestEffortCharsetEncoder.new
 
+    # Escapes any parenthesis in a string that are unescaped this uses
+    # a Ruby 1.9.1 regexp feature of negative look behind
+    def Utilities.escape_paren( str )
+      re = /(?<!\\)([\(\)])/          # Only match unescaped parens
+      str.gsub(re) { |s| '\\' + s }
     end
 
-    # Test String#encode works correctly with line endings.
-    # Some versions of Ruby (e.g. MRI <1.9, JRuby, Rubinius) line ending
-    # normalization does not work correctly or did not have #encode.
-    if ("\r".encode(:universal_newline => true) rescue nil) == LF &&
-      (LF.encode(:crlf_newline => true) rescue nil) == CRLF
-      # Using String#encode is better performing than Regexp
+    def Utilities.paren( str )
+      str = ::Mail::Utilities.unparen( str )
+      str = escape_paren( str )
+      '(' + str + ')'
+    end
 
-      def self.to_lf(input)
-        input.kind_of?(String) ? input.to_str.encode(input.encoding, :universal_newline => true) : ''
-      end
+    def Utilities.escape_bracket( str )
+      re = /(?<!\\)([\<\>])/          # Only match unescaped brackets
+      str.gsub(re) { |s| '\\' + s }
+    end
+
+    def Utilities.bracket( str )
+      str = ::Mail::Utilities.unbracket( str )
+      str = escape_bracket( str )
+      '<' + str + '>'
+    end
 
-      def self.to_crlf(input)
-        input.kind_of?(String) ? input.to_str.encode(input.encoding, :universal_newline => true).encode!(input.encoding, :crlf_newline => true) : ''
+    def Utilities.decode_base64(str)
+      if !str.end_with?("=") && str.length % 4 != 0
+        str = str.ljust((str.length + 3) & ~3, "=")
       end
+      str.unpack( 'm' ).first
+    end
+
+    def Utilities.encode_base64(str)
+      [str].pack( 'm' )
+    end
+
+    def Utilities.has_constant?(klass, string)
+      klass.const_defined?( string, false )
+    end
 
-    else
+    def Utilities.get_constant(klass, string)
+      klass.const_get( string )
+    end
+
+    def Utilities.transcode_charset(str, from_encoding, to_encoding = Encoding::UTF_8)
+      to_encoding = Encoding.find(to_encoding)
+      replacement_char = to_encoding == Encoding::UTF_8 ? '�' : '?'
+      charset_encoder.encode(str.dup, from_encoding).encode(to_encoding, :undef => :replace, :invalid => :replace, :replace => replacement_char)
+    end
+
+    # From Ruby stdlib Net::IMAP
+    def Utilities.encode_utf7(string)
+      string.gsub(/(&)|[^\x20-\x7e]+/) do
+        if $1
+          "&-"
+        else
+          base64 = [$&.encode(Encoding::UTF_16BE)].pack("m0")
+          "&" + base64.delete("=").tr("/", ",") + "-"
+        end
+      end.force_encoding(Encoding::ASCII_8BIT)
+    end
 
-      def self.to_lf(input)
-        input.kind_of?(String) ? input.to_str.gsub(/\r\n|\r/, LF) : ''
+    def Utilities.decode_utf7(utf7)
+      utf7.gsub(/&([^-]+)?-/n) do
+        if $1
+          ($1.tr(",", "/") + "===").unpack("m")[0].encode(Encoding::UTF_8, Encoding::UTF_16BE)
+        else
+          "&"
+        end
       end
+    end
 
-      if RUBY_VERSION >= '1.9'
-        # This 1.9 only regex can save a reasonable amount of time (~20%)
-        # by not matching "\r\n" so the string is returned unchanged in
-        # the common case.
-        CRLF_REGEX = Regexp.new("(?<!\r)\n|\r(?!\n)")
-      else
-        CRLF_REGEX = /\n|\r\n|\r/
+    def Utilities.b_value_encode(str, encoding = nil)
+      encoding = str.encoding.to_s
+      [Utilities.encode_base64(str), encoding]
+    end
+
+    def Utilities.b_value_decode(str)
+      match = str.match(/\=\?(.+)?\?[Bb]\?(.*)\?\=/m)
+      if match
+        charset = match[1]
+        str = Utilities.decode_base64(match[2])
+        str = charset_encoder.encode(str, charset)
       end
+      transcode_to_scrubbed_utf8(str)
+    rescue Encoding::UndefinedConversionError, ArgumentError, Encoding::ConverterNotFoundError, Encoding::InvalidByteSequenceError
+      warn "Encoding conversion failed #{$!}"
+      str.dup.force_encoding(Encoding::UTF_8)
+    end
+
+    def Utilities.q_value_encode(str, encoding = nil)
+      encoding = str.encoding.to_s
+      [Encodings::QuotedPrintable.encode(str), encoding]
+    end
 
-      def self.to_crlf(input)
-        input.kind_of?(String) ? input.to_str.gsub(CRLF_REGEX, CRLF) : ''
+    def Utilities.q_value_decode(str)
+      match = str.match(/\=\?(.+)?\?[Qq]\?(.*)\?\=/m)
+      if match
+        charset = match[1]
+        string = match[2].gsub(/_/, '=20')
+        # Remove trailing = if it exists in a Q encoding
+        string = string.sub(/\=$/, '')
+        str = Encodings::QuotedPrintable.decode(string)
+        str = charset_encoder.encode(str, charset)
+        # We assume that binary strings hold utf-8 directly to work around
+        # jruby/jruby#829 which subtly changes String#encode semantics.
+        str.force_encoding(Encoding::UTF_8) if str.encoding == Encoding::ASCII_8BIT
       end
+      transcode_to_scrubbed_utf8(str)
+    rescue Encoding::UndefinedConversionError, ArgumentError, Encoding::ConverterNotFoundError
+      warn "Encoding conversion failed #{$!}"
+      str.dup.force_encoding(Encoding::UTF_8)
+    end
 
+    def Utilities.param_decode(str, encoding)
+      str = uri_parser.unescape(str)
+      str = charset_encoder.encode(str, encoding) if encoding
+      transcode_to_scrubbed_utf8(str)
+    rescue Encoding::UndefinedConversionError, ArgumentError, Encoding::ConverterNotFoundError
+      warn "Encoding conversion failed #{$!}"
+      str.dup.force_encoding(Encoding::UTF_8)
     end
 
-    # Returns true if the object is considered blank.
-    # A blank includes things like '', '   ', nil,
-    # and arrays and hashes that have nothing in them.
+    def Utilities.param_encode(str)
+      encoding = str.encoding.to_s.downcase
+      language = Configuration.instance.param_encode_language
+      "#{encoding}'#{language}'#{uri_parser.escape(str)}"
+    end
+
+    def Utilities.uri_parser
+      URI::DEFAULT_PARSER
+    end
+
+    # Pick a Ruby encoding corresponding to the message charset. Most
+    # charsets have a Ruby encoding, but some need manual aliasing here.
     #
-    # This logic is mostly shared with ActiveSupport's blank?
-    def self.blank?(value)
-      if value.kind_of?(NilClass)
-        true
-      elsif value.kind_of?(String)
-        value !~ /\S/
+    # TODO: add this as a test somewhere:
+    #   Encoding.list.map { |e| [e.to_s.upcase == pick_encoding(e.to_s.downcase.gsub("-", "")), e.to_s] }.select {|a,b| !b}
+    #   Encoding.list.map { |e| [e.to_s == pick_encoding(e.to_s), e.to_s] }.select {|a,b| !b}
+    def Utilities.pick_encoding(charset)
+      charset = charset.to_s
+      encoding = case charset.downcase
+
+      # ISO-8859-8-I etc. http://en.wikipedia.org/wiki/ISO-8859-8-I
+      when /^iso[-_]?8859-(\d+)(-i)?$/
+        "ISO-8859-#{$1}"
+
+      # ISO-8859-15, ISO-2022-JP and alike
+      when /^iso[-_]?(\d{4})-?(\w{1,2})$/
+        "ISO-#{$1}-#{$2}"
+
+      # "ISO-2022-JP-KDDI"  and alike
+      when /^iso[-_]?(\d{4})-?(\w{1,2})-?(\w*)$/
+        "ISO-#{$1}-#{$2}-#{$3}"
+
+      # UTF-8, UTF-32BE and alike
+      when /^utf[\-_]?(\d{1,2})?(\w{1,2})$/
+        "UTF-#{$1}#{$2}".gsub(/\A(UTF-(?:16|32))\z/, '\\1BE')
+
+      # Windows-1252 and alike
+      when /^windows-?(.*)$/
+        "Windows-#{$1}"
+
+      when '8bit'
+        Encoding::ASCII_8BIT
+
+      # alternatives/misspellings of us-ascii seen in the wild
+      when /^iso[-_]?646(-us)?$/, 'us=ascii'
+        Encoding::ASCII
+
+      # Microsoft-specific alias for MACROMAN
+      when 'macintosh'
+        Encoding::MACROMAN
+
+      # Microsoft-specific alias for CP949 (Korean)
+      when 'ks_c_5601-1987'
+        Encoding::CP949
+
+      # Wrongly written Shift_JIS (Japanese)
+      when 'shift-jis'
+        Encoding::Shift_JIS
+
+      # GB2312 (Chinese charset) is a subset of GB18030 (its replacement)
+      when 'gb2312'
+        Encoding::GB18030
+
+      when 'cp-850'
+        Encoding::CP850
+
+      when 'latin2'
+        Encoding::ISO_8859_2
+
       else
-        value.respond_to?(:empty?) ? value.empty? : !value
+        charset
       end
+
+      convert_to_encoding(encoding)
+    end
+
+    def Utilities.string_byteslice(str, *args)
+      str.byteslice(*args)
     end
 
+    class << self
+      private
+
+      def convert_to_encoding(encoding)
+        if encoding.is_a?(Encoding)
+          encoding
+        else
+          # Fall back to ASCII for charsets that Ruby doesn't recognize
+          begin
+            Encoding.find(encoding)
+          rescue ArgumentError
+            Encoding::BINARY
+          end
+        end
+      end
+
+      def transcode_to_scrubbed_utf8(str)
+        decoded = str.encode(Encoding::UTF_8, :undef => :replace, :invalid => :replace, :replace => "�")
+        decoded.valid_encoding? ? decoded : decoded.encode(Encoding::UTF_16LE, :invalid => :replace, :replace => "�").encode(Encoding::UTF_8)
+      end
+    end
   end
 end

data/lib/mail/version.rb

@@ -3,8 +3,8 @@ module Mail
   module VERSION
 
     MAJOR = 2
-    MINOR = 6
-    PATCH = 6
+    MINOR = 8
+    PATCH = 1
     BUILD = nil
 
     STRING = [MAJOR, MINOR, PATCH, BUILD].compact.join('.')

data/lib/mail/yaml.rb

@@ -0,0 +1,30 @@
+require 'yaml'
+
+module Mail
+  module YAML
+    def self.load(yaml)
+      permitted_classes = [
+        Symbol,
+
+        Mail::Body,
+
+        # Delivery methods as listed in mail/configuration.rb
+        Mail::SMTP,
+        Mail::Sendmail,
+        Mail::Exim,
+        Mail::FileDelivery,
+        Mail::SMTPConnection,
+        Mail::TestMailer,
+        Mail::LoggerDelivery,
+
+        Mail.delivery_method.class,
+      ]
+
+      if Gem::Version.new(Psych::VERSION) >= Gem::Version.new('3.1.0.pre1')
+        ::YAML.safe_load(yaml, :permitted_classes => permitted_classes)
+      else
+        ::YAML.safe_load(yaml, permitted_classes)
+      end
+    end
+  end
+end

data/lib/mail.rb

@@ -7,44 +7,13 @@ module Mail # :doc:
 
   require 'uri'
   require 'net/smtp'
-
-  begin
-    # Use mime/types/columnar if available, for reduced memory usage
-    require 'mime/types/columnar'
-  rescue LoadError
-    require 'mime/types'
-  end
-
-  if RUBY_VERSION <= '1.8.6'
-    begin
-      require 'tlsmail'
-    rescue LoadError
-      raise "You need to install tlsmail if you are using ruby <= 1.8.6"
-    end
-  end
-
-  if RUBY_VERSION >= "1.9.0"
-    require 'mail/version_specific/ruby_1_9'
-    RubyVer = Ruby19
-  else
-    require 'mail/version_specific/ruby_1_8'
-    RubyVer = Ruby18
-  end
+  require 'mini_mime'
 
   require 'mail/version'
 
-  require 'mail/core_extensions/string'
-  require 'mail/core_extensions/smtp' if RUBY_VERSION < '1.9.3'
   require 'mail/indifferent_hash'
 
-  # Only load our multibyte extensions if AS is not already loaded
-  if defined?(ActiveSupport)
-    require 'active_support/inflector'
-  else
-    require 'mail/core_extensions/string/access'
-    require 'mail/core_extensions/string/multibyte'
-    require 'mail/multibyte'
-  end
+  require 'mail/multibyte'
 
   require 'mail/constants'
   require 'mail/utilities'
@@ -81,8 +50,6 @@ module Mail # :doc:
 
   require 'mail/envelope'
 
-  register_autoload :Parsers, "mail/parsers"
-
   # Autoload header field elements and transfer encodings.
   require 'mail/elements'
   require 'mail/encodings'
@@ -93,6 +60,9 @@ module Mail # :doc:
   require 'mail/matchers/has_sent_mail'
   require 'mail/matchers/attachment_matchers.rb'
 
+  # Deprecated will be removed in 3.0 release
+  require 'mail/check_delivery_params'
+
   # Finally... require all the Mail.methods
   require 'mail/mail'
 end