Class: Mail::Ruby19

Inherits:
Object
  • Object
show all
Defined in:
lib/mail/version_specific/ruby_1_9.rb

Defined Under Namespace

Classes: BestEffortCharsetEncoder, StrictCharsetEncoder

Class Attribute Summary collapse

Class Method Summary collapse

Class Attribute Details

.charset_encoderObject

Returns the value of attribute charset_encoder



32
33
34
# File 'lib/mail/version_specific/ruby_1_9.rb', line 32

def charset_encoder
  @charset_encoder
end

Class Method Details

.b_value_decode(str) ⇒ Object



85
86
87
88
89
90
91
92
93
94
95
96
97
# File 'lib/mail/version_specific/ruby_1_9.rb', line 85

def Ruby19.b_value_decode(str)
  match = str.match(/\=\?(.+)?\?[Bb]\?(.*)\?\=/m)
  if match
    charset = match[1]
    str = Ruby19.decode_base64(match[2])
    str = charset_encoder.encode(str, charset)
  end
  decoded = str.encode(Encoding::UTF_8, :invalid => :replace, :replace => "")
  decoded.valid_encoding? ? decoded : decoded.encode(Encoding::UTF_16LE, :invalid => :replace, :replace => "").encode(Encoding::UTF_8)
rescue Encoding::UndefinedConversionError, ArgumentError, Encoding::ConverterNotFoundError
  warn "Encoding conversion failed #{$!}"
  str.dup.force_encoding(Encoding::UTF_8)
end

.b_value_encode(str, encoding = nil) ⇒ Object



80
81
82
83
# File 'lib/mail/version_specific/ruby_1_9.rb', line 80

def Ruby19.b_value_encode(str, encoding = nil)
  encoding = str.encoding.to_s
  [Ruby19.encode_base64(str), encoding]
end

.bracket(str) ⇒ Object



54
55
56
57
58
# File 'lib/mail/version_specific/ruby_1_9.rb', line 54

def Ruby19.bracket( str )
  str = $1 if str =~ /^\<(.*)?\>$/
  str = escape_bracket( str )
  '<' + str + '>'
end

.decode_base64(str) ⇒ Object



60
61
62
# File 'lib/mail/version_specific/ruby_1_9.rb', line 60

def Ruby19.decode_base64(str)
  str.unpack( 'm' ).first
end

.encode_base64(str) ⇒ Object



64
65
66
# File 'lib/mail/version_specific/ruby_1_9.rb', line 64

def Ruby19.encode_base64(str)
  [str].pack( 'm' )
end

.escape_bracket(str) ⇒ Object



49
50
51
52
# File 'lib/mail/version_specific/ruby_1_9.rb', line 49

def Ruby19.escape_bracket( str )
  re = /(?<!\\)([\<\>])/          # Only match unescaped brackets
  str.gsub(re) { |s| '\\' + s }
end

.escape_paren(str) ⇒ Object

Escapes any parenthesis in a string that are unescaped this uses a Ruby 1.9.1 regexp feature of negative look behind



38
39
40
41
# File 'lib/mail/version_specific/ruby_1_9.rb', line 38

def Ruby19.escape_paren( str )
  re = /(?<!\\)([\(\)])/          # Only match unescaped parens
  str.gsub(re) { |s| '\\' + s }
end

.get_constant(klass, string) ⇒ Object



72
73
74
# File 'lib/mail/version_specific/ruby_1_9.rb', line 72

def Ruby19.get_constant(klass, string)
  klass.const_get( string )
end

.has_constant?(klass, string) ⇒ Boolean



68
69
70
# File 'lib/mail/version_specific/ruby_1_9.rb', line 68

def Ruby19.has_constant?(klass, string)
  klass.const_defined?( string, false )
end

.param_decode(str, encoding) ⇒ Object



124
125
126
127
128
# File 'lib/mail/version_specific/ruby_1_9.rb', line 124

def Ruby19.param_decode(str, encoding)
  str = uri_parser.unescape(str)
  str = charset_encoder.encode(str, encoding) if encoding
  str
end

.param_encode(str) ⇒ Object



130
131
132
133
134
# File 'lib/mail/version_specific/ruby_1_9.rb', line 130

def Ruby19.param_encode(str)
  encoding = str.encoding.to_s.downcase
  language = Configuration.instance.param_encode_language
  "#{encoding}'#{language}'#{uri_parser.escape(str)}"
end

.paren(str) ⇒ Object



43
44
45
46
47
# File 'lib/mail/version_specific/ruby_1_9.rb', line 43

def Ruby19.paren( str )
  str = $1 if str =~ /^\((.*)?\)$/
  str = escape_paren( str )
  '(' + str + ')'
end

.pick_encoding(charset) ⇒ Object

Pick a Ruby encoding corresponding to the message charset. Most charsets have a Ruby encoding, but some need manual aliasing here.

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}


146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
# File 'lib/mail/version_specific/ruby_1_9.rb', line 146

def Ruby19.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
    charset
  end

  convert_to_encoding(encoding)
end

.q_value_decode(str) ⇒ Object



104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
# File 'lib/mail/version_specific/ruby_1_9.rb', line 104

def Ruby19.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(/\=$/, '')
    str = Encodings::QuotedPrintable.decode(string)
    str = charset_encoder.encode(str, charset)
    # We assume that binary strings hold utf-8 directly to work around
    # jruby/jruby#829 which subtly changes String#encode semantics.
    str.force_encoding(Encoding::UTF_8) if str.encoding == Encoding::ASCII_8BIT
  end
  decoded = str.encode(Encoding::UTF_8, :invalid => :replace, :replace => "")
  decoded.valid_encoding? ? decoded : decoded.encode(Encoding::UTF_16LE, :invalid => :replace, :replace => "").encode(Encoding::UTF_8)
rescue Encoding::UndefinedConversionError, ArgumentError, Encoding::ConverterNotFoundError
  warn "Encoding conversion failed #{$!}"
  str.dup.force_encoding(Encoding::UTF_8)
end

.q_value_encode(str, encoding = nil) ⇒ Object



99
100
101
102
# File 'lib/mail/version_specific/ruby_1_9.rb', line 99

def Ruby19.q_value_encode(str, encoding = nil)
  encoding = str.encoding.to_s
  [Encodings::QuotedPrintable.encode(str), encoding]
end

.transcode_charset(str, from_encoding, to_encoding = Encoding::UTF_8) ⇒ Object



76
77
78
# File 'lib/mail/version_specific/ruby_1_9.rb', line 76

def Ruby19.transcode_charset(str, from_encoding, to_encoding = Encoding::UTF_8)
  charset_encoder.encode(str.dup, from_encoding).encode(to_encoding, :undef => :replace, :invalid => :replace, :replace => '')
end

.uri_parserObject



136
137
138
# File 'lib/mail/version_specific/ruby_1_9.rb', line 136

def Ruby19.uri_parser
  @uri_parser ||= URI::Parser.new
end