mail 2.6.6 → 2.7.1

data/lib/mail/parsers/received_parser.rl

@@ -0,0 +1,84 @@
+# frozen_string_literal: true
+require 'mail/utilities'
+require 'mail/parser_tools'
+
+%%{
+  machine date_time;
+  alphtype int;
+
+  # Received Tokens
+  action received_tokens_s { received_tokens_s = p }
+  action received_tokens_e { received.info = chars(data, received_tokens_s, p-1) }
+
+  # Date
+  action date_s { date_s = p }
+  action date_e { received.date = chars(data, date_s, p-1).strip }
+
+  # Time
+  action time_s { time_s = p }
+  action time_e { received.time = chars(data, time_s, p-1) }
+
+  # No-op actions
+  action address_s {}
+  action address_e {}
+  action angle_addr_s {}
+  action ctime_date_s {}
+  action ctime_date_e {}
+  action comment_e {}
+  action comment_s {}
+  action phrase_s {}
+  action phrase_e {}
+  action domain_e {}
+  action domain_s {}
+  action local_dot_atom_e {}
+  action local_dot_atom_pre_comment_e {}
+  action local_dot_atom_pre_comment_s {}
+  action local_dot_atom_s {}
+  action qstr_e {}
+  action qstr_s {}
+  action local_quoted_string_s {}
+  action local_quoted_string_e {}
+  action obs_domain_list_s {}
+  action obs_domain_list_e {}
+  action group_name_s {}
+  action group_name_e {}
+  action msg_id_s {}
+  action msg_id_e {}
+
+  include rfc5322 "rfc5322.rl";
+  main := received;
+}%%
+
+module Mail::Parsers
+  module ReceivedParser
+    extend Mail::ParserTools
+
+    ReceivedStruct = Struct.new(:date, :time, :info, :error)
+
+    %%write data noprefix;
+
+    def self.parse(data)
+      data = data.dup.force_encoding(Encoding::ASCII_8BIT) if data.respond_to?(:force_encoding)
+
+      raise Mail::Field::NilParseError.new(Mail::ReceivedElement) if data.nil?
+
+      # Parser state
+      received = ReceivedStruct.new
+      received_tokens_s = date_s = time_s = nil
+
+      # 5.1 Variables Used by Ragel
+      p = 0
+      eof = pe = data.length
+      stack = []
+
+      %%write init;
+      %%write exec;
+
+      if p != eof || cs < %%{ write first_final; }%%
+        raise Mail::Field::IncompleteParseError.new(Mail::ReceivedElement, data, p)
+      end
+
+      received
+    end
+  end
+end

data/lib/mail/parsers/rfc2045_content_transfer_encoding.rl

@@ -0,0 +1,13 @@
+%%{
+  # RFC 2045 Section 6.1: Content-Transfer-Encoding Header Field
+  # https://tools.ietf.org/html/rfc2045#section-6.1
+  machine rfc2045_content_transfer_encoding;
+  alphtype int;
+
+  include rfc2045_content_type "rfc2045_content_type.rl";
+
+  encoding = ('7bits' | '8bits' | '7bit' | '8bit' | 'binary' |
+              'quoted-printable' | 'base64' | ietf_token |
+              custom_x_token) >encoding_s %encoding_e;
+  content_transfer_encoding = CFWS? encoding CFWS? ";"? CFWS?;
+}%%

data/lib/mail/parsers/rfc2045_content_type.rl

@@ -0,0 +1,25 @@
+%%{
+  # RFC 2045 Section 5.1: Content-Type Header Field
+  # https://tools.ietf.org/html/rfc2045#section-5.1
+  # Previously: https://tools.ietf.org/html/rfc1049#section-3
+  machine rfc2045_content_type;
+  alphtype int;
+
+  include rfc5322_lexical_tokens "rfc5322_lexical_tokens.rl";
+
+  token = 0x21..0x27 | 0x2a..0x2b | 0x2c..0x2e | 0x30..0x39 | 0x41..0x5a | 0x5e..0x7e;
+  value = (quoted_string | (token -- '"' | 0x3d)+) >param_val_s %param_val_e;
+  attribute = (token+) >param_attr_s %param_attr_e;
+  parameter = CFWS? attribute "=" value CFWS?;
+
+  ietf_token = token+;
+  custom_x_token = 'x'i "-" token+;
+  extension_token = ietf_token | custom_x_token;
+  discrete_type = 'text'i | 'image'i | 'audio'i | 'video'i |
+                  'application'i | extension_token;
+  composite_type = 'message'i | 'multipart'i | extension_token;
+  iana_token = token+;
+  main_type = (discrete_type | composite_type) >main_type_s %main_type_e;
+  sub_type = (extension_token | iana_token) >sub_type_s %sub_type_e;
+  content_type = main_type "/" sub_type (((CFWS? ";"+) | CFWS) parameter CFWS?)*;
+}%%

data/lib/mail/parsers/rfc2045_mime.rl

@@ -0,0 +1,16 @@
+%%{
+  # RFC 2045 MIME
+  # https://tools.ietf.org/html/rfc2045
+  machine rfc2045_mime;
+  alphtype int;
+
+  include rfc5322_lexical_tokens "rfc5322_lexical_tokens.rl";
+
+  # 4. MIME-Version Header Field
+  # https://tools.ietf.org/html/rfc2045#section-4
+  mime_version = CFWS?
+            (DIGIT+ >major_digits_s %major_digits_e)
+            comment? "." comment?
+            (DIGIT+ >minor_digits_s %minor_digits_e)
+            CFWS?;
+}%%

data/lib/mail/parsers/rfc2183_content_disposition.rl

@@ -0,0 +1,15 @@
+%%{
+  # RFC 2183 The Content-Disposition Header Field
+  # https://tools.ietf.org/html/rfc2183#section-2
+  #
+  # TODO: recognize filename, size, creation date, etc.
+  machine rfc2183_content_disposition;
+  alphtype int;
+
+  include rfc2045_content_type "rfc2045_content_type.rl";
+
+  disposition_type = 'inline'i | 'attachment'i | extension_token;
+  disposition_parm = parameter;
+  disposition = (disposition_type >disp_type_s %disp_type_e)
+                (";" disposition_parm)*;
+}%%

data/lib/mail/parsers/rfc3629_utf8.rl

@@ -0,0 +1,19 @@
+%%{
+  # RFC 3629 4. Syntax of UTF-8 Byte Sequences
+  # https://tools.ietf.org/html/rfc3629#section-4
+  machine rfc3629_utf8;
+  alphtype int;
+
+  utf8_tail  = 0x80..0xBF;
+
+  utf8_2byte = 0xC2..0xDF utf8_tail;
+  utf8_3byte = 0xE0       0xA0..0xBF  utf8_tail |
+               0xE1..0xEC utf8_tail   utf8_tail |
+               0xED       0x80..0x9F  utf8_tail |
+               0xEE..0xEF utf8_tail   utf8_tail;
+  utf8_4byte = 0xF0       0x90..0xBF  utf8_tail utf8_tail |
+               0xF1..0xF3 utf8_tail   utf8_tail utf8_tail |
+               0xF4       0x80..0x8F  utf8_tail utf8_tail;
+
+  utf8_non_ascii = utf8_2byte | utf8_3byte | utf8_4byte;
+}%%

data/lib/mail/parsers/rfc5234_abnf_core_rules.rl

@@ -0,0 +1,22 @@
+%%{
+  # RFC 5234 B.1. Core Rules
+  # https://tools.ietf.org/html/rfc5234#appendix-B.1
+  machine rfc5234_abnf_core_rules;
+  alphtype int;
+
+  include rfc3629_utf8 "rfc3629_utf8.rl";
+
+  LF = "\n";
+  CR = "\r";
+  CRLF = "\r\n";
+  SP = " ";
+  HTAB = "\t";
+  WSP = SP | HTAB;
+  DQUOTE = '"';
+  DIGIT = [0-9];
+  ALPHA = [a-zA-Z];
+
+  # RFC6532 extension for UTF-8 content
+  rfc5234_VCHAR = 0x21..0x7e;
+  VCHAR = rfc5234_VCHAR | utf8_non_ascii;
+}%%

data/lib/mail/parsers/rfc5322.rl

@@ -0,0 +1,59 @@
+%%{
+  # RFC 5322 Internet Message Format
+  # https://tools.ietf.org/html/rfc5322
+  #
+  # RFC 6854 Update to Internet Message Format to Allow Group Syntax in the "From:" and "Sender:" Header Fields
+  # https://tools.ietf.org/html/rfc6854
+  machine rfc5322;
+  alphtype int;
+
+  include rfc5234_abnf_core_rules "rfc5234_abnf_core_rules.rl";
+
+  # 3.2. Lexical Tokens
+  include rfc5322_lexical_tokens "rfc5322_lexical_tokens.rl";
+
+  # 3.3. Date and Time Specification
+  include rfc5322_date_time "rfc5322_date_time.rl";
+
+  # 3.4. Address Specification
+  include rfc5322_address "rfc5322_address.rl";
+
+  # 3.5. Overall Message Syntax
+  #rfc5322_text = 0x01..0x09 | "\v" | "\f" | 0x0e..0x1f;
+  #text = rfc5322_text | utf8_non_ascii; # RFC6532 for UTF-8
+  #obs_body = ((LF* CR* ((0x00 | text) LF* CR*)*) | CRLF)*
+  #body = ((text{,998} CRLF)* text{,998}) | obs_body;
+  #message = (fields | obs_fields) (CRLF body)?;
+
+
+  # 3.6. Field Definitions
+
+  # 3.6.4. Identification Fields
+  obs_id_left = local_part;
+  id_left = dot_atom_text | obs_id_left;
+  # id_right modifications to support multiple '@' in msg_id.
+  msg_id_atext = ALPHA | DIGIT | "!" | "#" | "$" | "%" | "&" | "'" | "*" |
+                 "+" | "-" | "/" | "=" | "?" | "^" | "_" | "`" | "{" | "|" |
+                 "}" | "~" | "@";
+  msg_id_dot_atom_text = (msg_id_atext+ "."?)+;
+  obs_id_right = domain;
+  no_fold_literal = "[" (dtext)* "]";
+  id_right = msg_id_dot_atom_text | no_fold_literal | obs_id_right;
+  msg_id = (CFWS)?
+           (("<" id_left "@" id_right ">") >msg_id_s %msg_id_e)
+           (CFWS)?;
+  message_ids = msg_id (CFWS? msg_id)*;
+
+
+  # 3.6.7 Trace Fields
+  # Added CFWS? to increase robustness (qmail likes to include a comment)
+  received_token = word | angle_addr | addr_spec_no_angle_brackets | domain;
+  received = ((CFWS? received_token*) >received_tokens_s %received_tokens_e)
+              ";" date_time;
+
+  # Envelope From
+  ctime_date = day_name " "+ month " "+ day " " time_of_day " " year;
+  null_sender = ('<>' ' '{0,1});
+  envelope_from = (addr_spec_no_angle_brackets | null_sender) >address_s %address_e " "
+                  (ctime_date >ctime_date_s %ctime_date_e);
+}%%

data/lib/mail/parsers/rfc5322_address.rl

@@ -0,0 +1,72 @@
+%%{
+  # RFC 5322 Internet Message Format
+  # Section 3.4. Address Specification
+  # https://tools.ietf.org/html/rfc5322#section-3.4
+  machine rfc5322_address;
+  alphtype int;
+
+  include rfc5234_abnf_core_rules "rfc5234_abnf_core_rules.rl";
+  include rfc5322_lexical_tokens "rfc5322_lexical_tokens.rl";
+
+  # local_part:
+  domain_text = (DQUOTE (FWS? qcontent)+ FWS? DQUOTE) | atext+;
+  local_dot_atom_text = ("."* domain_text "."*)+;
+  local_dot_atom = CFWS?
+                   (local_dot_atom_text >local_dot_atom_s %local_dot_atom_pre_comment_e)
+                   CFWS?;
+  obs_local_part = word ("." word)*;
+  local_part = (local_dot_atom >local_dot_atom_s %local_dot_atom_e |
+                (quoted_string %local_quoted_string_e) |
+                obs_local_part);
+
+  # Treetop parser behavior was to ignore addresses missing '@' inside of angle
+  # brackets. This construction preserves that behavior.
+  local_part_no_capture = (local_dot_atom | quoted_string | obs_local_part);
+
+  # domain:
+  domain_dot_atom_text = "."* domain_text ("."* domain_text)*;
+  obs_dtext = obs_NO_WS_CTL | quoted_pair;
+  rfc5322_dtext = 0x21..0x5a | 0x5e..0x7e | obs_dtext;
+  dtext = rfc5322_dtext | utf8_non_ascii; # RFC6532 for UTF-8
+  domain_dot_atom = CFWS? domain_dot_atom_text (CFWS? >(comment_after_address,1));
+  domain_literal = CFWS? "[" (FWS? dtext)* FWS? "]" CFWS?;
+  obs_domain = atom ("." atom)*;
+  domain = (domain_dot_atom | domain_literal | obs_domain) >domain_s %domain_e;
+
+  # 3.4.1. Addr-Spec Specification
+
+  # The %(end_addr,N) priority resolves uncertainty when whitespace
+  # after an addr_spec could cause it to be interpreted as a
+  # display name: "bar@example.com ,..."
+
+  addr_spec_in_angle_brackets =
+    (local_part "@" domain) %(end_addr,1) |
+    local_part_no_capture   %(end_addr,0);
+
+  addr_spec_no_angle_brackets =
+    (local_part "@" domain) %(end_addr,1) |
+    local_part              %(end_addr,0);
+
+  # angle_addr:
+  obs_domain_list = (CFWS | ",")* "@" domain ("," CFWS? ("@" domain)?)*;
+  obs_route = (obs_domain_list ":") >obs_domain_list_s %obs_domain_list_e;
+  obs_angle_addr = CFWS? "<" obs_route? addr_spec_in_angle_brackets ">" CFWS?;
+
+  angle_addr = CFWS? ("<" >angle_addr_s) addr_spec_in_angle_brackets ">" CFWS? |
+                obs_angle_addr;
+
+  # 3.4. Address Specification
+  display_name = phrase;
+  name_addr = display_name? %(end_addr,2) angle_addr;
+  mailbox = (name_addr | addr_spec_no_angle_brackets) >address_s %address_e;
+  obs_mbox_list = (CFWS? ",")* mailbox ("," (mailbox | CFWS)?)*;
+  mailbox_list = (mailbox (("," | ";") mailbox)*) | obs_mbox_list;
+  obs_group_list = (CFWS? ",")+ CFWS?;
+  group_list = mailbox_list | CFWS | obs_group_list;
+  group = (display_name >group_name_s %group_name_e) ":"
+            (group_list?) ";" CFWS?;
+  address = group | mailbox;
+  #obs_addr_list = (CFWS? ",")* address ("," (address | CFWS)?)*;
+  address_lists = address? %(comment_after_address,0)
+                  (FWS* ("," | ";") FWS* address?)*;
+}%%

data/lib/mail/parsers/{ragel/date_time.rl → rfc5322_date_time.rl}

@@ -1,6 +1,11 @@
 %%{
+  # RFC 5322 Internet Message Format
+  # Section 3.3. Date and Time Specification
+  # https://tools.ietf.org/html/rfc5322#section-3.3
+  machine rfc5322_date_time;
+  alphtype int;
 
-  machine date_time;
+  include rfc5322_lexical_tokens "rfc5322_lexical_tokens.rl";
 
   # day_of_week
   day_name = "Mon" | "Tue" | "Wed" | "Thu" | "Fri" | "Sat" | "Sun";
@@ -27,4 +32,6 @@
   zone = FWS ((("+" | "-") DIGIT DIGIT DIGIT DIGIT) | obs_zone);
   time = time_of_day zone;
 
+  date_time = (day_of_week ",")?
+              (date >date_s %date_e) <: (time >time_s %time_e) CFWS?;
 }%%

data/lib/mail/parsers/rfc5322_lexical_tokens.rl

@@ -0,0 +1,60 @@
+%%{
+  # RFC 5322 Internet Message Format
+  # Section 3.2. Lexical Tokens
+  # https://tools.ietf.org/html/rfc5322#section-3.2
+  machine rfc5322_lexical_tokens;
+  alphtype int;
+
+  include rfc5234_abnf_core_rules "rfc5234_abnf_core_rules.rl";
+
+  # 3.2.1.  Quoted characters
+  obs_NO_WS_CTL = 0x01..0x08 | "\v" | "\f" | 0x0e..0x1f | 0x7f;
+  obs_qp = "\\" (0x00 | obs_NO_WS_CTL | LF | CR);
+  quoted_pair = ("\\" (VCHAR | WSP)) | obs_qp;
+
+  # 3.2.2. Folding White Space and Comments
+  obs_FWS = (CRLF? WSP)+;
+  FWS = (WSP* CRLF WSP+) | (CRLF WSP+) | obs_FWS;
+
+  obs_ctext = obs_NO_WS_CTL;
+  rfc5322_ctext = 0x21..0x27 | 0x2a..0x5b | 0x5d..0x7e | obs_ctext;
+  ctext = rfc5322_ctext | utf8_non_ascii; # RFC6532 for UTF-8
+
+  # Recursive comments
+  action comment_begin { fcall comment_tail; }
+  action comment_exit { fret; }
+  ccontent = ctext | quoted_pair | "(" @comment_begin;
+  comment_tail := ((FWS? ccontent)* >comment_s) FWS? ")" @comment_exit;
+  comment = "(" @comment_begin %comment_e;
+  CFWS = ((FWS? comment)+ FWS?) | FWS;
+
+  # 3.2.3. Atom
+  rfc5322_atext = ALPHA | DIGIT | "!" | "#" | "$" | "%" | "&" |
+                  "'" | "*" | "+" | "-" | "/" | "=" | "?" | "^" |
+                  "_" | "`" | "{" | "|" | "}" | "~";
+  atext = rfc5322_atext | utf8_non_ascii; # RFC6532 for UTF-8
+  atom = CFWS? atext+ CFWS?;
+  dot_atom_text = atext ("." atext)*;
+  dot_atom = CFWS? dot_atom_text CFWS?;
+
+  # 3.2.4. Quoted Strings
+  obs_qtext = obs_NO_WS_CTL;
+  rfc5322_qtext = 0x21 | 0x23..0x5b | 0x5d..0x7e | obs_qtext;
+  qtext = rfc5322_qtext | utf8_non_ascii; # RFC6532 for UTF-8
+
+  qcontent = qtext | quoted_pair;
+  quoted_string = CFWS?
+                  (DQUOTE
+                    (((FWS? qcontent)* FWS?) >qstr_s %qstr_e)
+                  DQUOTE)
+                  CFWS?;
+
+  # 3.2.5. Miscellaneous Tokens
+  word = atom | quoted_string;
+
+  obs_phrase = (word | "." | "@")+;
+  phrase = (obs_phrase | word+) >phrase_s %phrase_e;
+
+  # Not part of RFC, used for keywords per 3.6.5 Information Fields
+  phrase_lists = phrase ("," FWS* phrase)*;
+}%%

data/lib/mail/parsers.rb

@@ -1,27 +1,19 @@
 # frozen_string_literal: true
-module Mail
-  module Parsers
+# Ragel-generated parsers are full of known warnings. Suppress them.
+begin
+  orig, $VERBOSE = $VERBOSE, nil
 
-    # Low-level ragel based parsers
-    require 'mail/parsers/ragel'
-
-    AddressListStruct = Struct.new(:addresses, :group_names, :error)
-    AddressStruct = Struct.new(:raw, :domain, :comments, :local,
-                             :obs_domain_list, :display_name, :group, :error)
-    ContentDispositionStruct = Struct.new(:disposition_type, :parameters, :error)
-    ContentLocationStruct = Struct.new(:location, :error)
-    ContentTransferEncodingStruct = Struct.new(:encoding, :error)
-    ContentTypeStruct = Struct.new(:main_type, :sub_type, :parameters, :error)
-    DateTimeStruct = Struct.new(:date_string, :time_string, :error)
-    EnvelopeFromStruct = Struct.new(:address, :ctime_date, :error)
-    MessageIdsStruct = Struct.new(:message_ids, :error)
-    MimeVersionStruct = Struct.new(:major, :minor, :error)
-    PhraseListsStruct = Struct.new(:phrases, :error)
-    ReceivedStruct = Struct.new(:date, :time, :info, :error)
-
-    require 'mail/parsers/ragel/parser_info'
-    Ragel::FIELD_PARSERS.each do |field_parser|
-      require "mail/parsers/#{field_parser}_parser"
-    end
-  end
+  require 'mail/parsers/address_lists_parser'
+  require 'mail/parsers/content_disposition_parser'
+  require 'mail/parsers/content_location_parser'
+  require 'mail/parsers/content_transfer_encoding_parser'
+  require 'mail/parsers/content_type_parser'
+  require 'mail/parsers/date_time_parser'
+  require 'mail/parsers/envelope_from_parser'
+  require 'mail/parsers/message_ids_parser'
+  require 'mail/parsers/mime_version_parser'
+  require 'mail/parsers/phrase_lists_parser'
+  require 'mail/parsers/received_parser'
+ensure
+  $VERBOSE = orig
 end

data/lib/mail/part.rb

@@ -21,7 +21,7 @@ module Mail
     
     def inline_content_id
       # TODO: Deprecated in 2.2.2 - Remove in 2.3
-      STDERR.puts("Part#inline_content_id is deprecated, please call Part#cid instead")
+      warn("Part#inline_content_id is deprecated, please call Part#cid instead")
       cid
     end
     
@@ -104,8 +104,8 @@ module Mail
     
     # A part may not have a header.... so, just init a body if no header
     def parse_message
-      header_part, body_part = raw_source.split(/#{CRLF}#{WSP}*#{CRLF}/m, 2)
-      if header_part =~ HEADER_LINE
+      header_part, body_part = raw_source.split(/#{Constants::CRLF}#{Constants::WSP}*#{Constants::CRLF}/m, 2)
+      if header_part =~ Constants::HEADER_LINE
         self.header = header_part
         self.body   = body_part
       else

data/lib/mail/parts_list.rb

@@ -55,7 +55,7 @@ module Mail
         # OK, 10000 is arbitrary... if anyone actually wants to explicitly sort 10000 parts of a
         # single email message... please show me a use case and I'll put more work into this method,
         # in the meantime, it works :)
-        [get_order_value(a, order), i += 1]
+        get_order_value(a, order) << (i += 1)
       end
       @parts.clear
       sorted.each { |p| @parts << p }
@@ -64,11 +64,10 @@ module Mail
   private
 
     def get_order_value(part, order)
-      if part.respond_to?(:content_type) && !part[:content_type].nil?
-        order.index(part[:content_type].string.downcase) || 10000
-      else
-        10000
-      end
+      is_attachment = part.respond_to?(:attachment?) && part.attachment?
+      has_content_type = part.respond_to?(:content_type) && !part[:content_type].nil?
+
+      [is_attachment ? 1 : 0, (has_content_type ? order.index(part[:content_type].string.downcase) : nil) || 10000]
     end
 
   end

data/lib/mail/utilities.rb

@@ -1,5 +1,7 @@
 # encoding: utf-8
 # frozen_string_literal: true
+require 'mail/constants'
+
 module Mail
   module Utilities
 
@@ -22,9 +24,9 @@ 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')
+        ascii_str = str.to_s.dup.force_encoding('ASCII-8BIT')
         if (PHRASE_UNSAFE === ascii_str)
           dquote(ascii_str).force_encoding(original_encoding)
         else
@@ -43,7 +45,17 @@ module Mail
     # 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,
@@ -77,6 +89,7 @@ module Mail
         str
       end
     end
+    module_function :unquote
 
     # Removes any \-escaping.
     #
@@ -192,7 +205,7 @@ 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)
     end
@@ -238,40 +251,59 @@ module Mail
 
     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 self.binary_unsafe_to_lf(string) #:nodoc:
+      string.gsub(/\r\n|\r/, LF)
+    end
 
-      def self.to_lf(input)
-        input.kind_of?(String) ? input.to_str.encode(input.encoding, :universal_newline => true) : ''
+    TO_CRLF_REGEX =
+      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.
+        Regexp.new("(?<!\r)\n|\r(?!\n)")
+      else
+        /\n|\r\n|\r/
       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) : ''
-      end
+    def self.binary_unsafe_to_crlf(string) #:nodoc:
+      string.gsub(TO_CRLF_REGEX, CRLF)
+    end
 
+    if RUBY_VERSION < '1.9'
+      def self.safe_for_line_ending_conversion?(string) #:nodoc:
+        string.ascii_only?
+      end
     else
-
-      def self.to_lf(input)
-        input.kind_of?(String) ? input.to_str.gsub(/\r\n|\r/, LF) : ''
+      def self.safe_for_line_ending_conversion?(string) #:nodoc:
+        if string.encoding == Encoding::BINARY
+          string.ascii_only?
+        else
+          string.valid_encoding?
+        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)")
+    # 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
-        CRLF_REGEX = /\n|\r\n|\r/
+        string
       end
+    end
 
-      def self.to_crlf(input)
-        input.kind_of?(String) ? input.to_str.gsub(CRLF_REGEX, CRLF) : ''
+    # 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
 
     # Returns true if the object is considered blank.
@@ -288,6 +320,5 @@ module Mail
         value.respond_to?(:empty?) ? value.empty? : !value
       end
     end
-
   end
 end

data/lib/mail/version.rb

@@ -3,8 +3,8 @@ module Mail
   module VERSION
 
     MAJOR = 2
-    MINOR = 6
-    PATCH = 6
+    MINOR = 7
+    PATCH = 1
     BUILD = nil
 
     STRING = [MAJOR, MINOR, PATCH, BUILD].compact.join('.')