Module: XSD::Charset

Defined in:
lib/xsd/charset.rb

Defined Under Namespace

Classes: CharsetConversionError, CharsetError, UnknownCharsetError, XSDError

Constant Summary collapse

EncodingConvertMap =

Maps

{}
CharsetMap =
{
  'NONE' => 'us-ascii',
  'EUC' => 'euc-jp',
  'SJIS' => 'shift_jis',
  'UTF8' => 'utf-8',
  'X_ISO_8859_1' => 'iso-8859-1',
  'X_UNKNOWN' => nil,
}
CharsetStrCache =
{}
USASCIIRegexp =

XML 1.0 restricted.

Regexp.new("\\A#{us_ascii}*\\z", nil, 'n')
EUCRegexp =
Regexp.new("\\A#{character_euc}*\\z", nil, 'n')
SJISRegexp =
Regexp.new("\\A#{character_sjis}*\\z", nil, 'n')
UTF8Regexp =
Regexp.new("\\A#{character_utf8}*\\z", nil, 'n')

Class Method Summary collapse

Class Method Details

.charset_label(encoding) ⇒ Object


119
120
121
# File 'lib/xsd/charset.rb', line 119

def Charset.charset_label(encoding)
  CharsetMap[encoding.upcase]
end

.charset_str(label) ⇒ Object


123
124
125
126
127
128
129
# File 'lib/xsd/charset.rb', line 123

def Charset.charset_str(label)
  if CharsetMap.respond_to?(:key)
    CharsetStrCache[label] ||= CharsetMap.key(label.downcase) || 'X_UNKNOWN'
  else
    CharsetStrCache[label] ||= CharsetMap.index(label.downcase) || 'X_UNKNOWN'
  end
end

.encodingObject

handlers


87
88
89
# File 'lib/xsd/charset.rb', line 87

def Charset.encoding
  @internal_encoding
end

.encoding=(encoding) ⇒ Object


91
92
93
94
# File 'lib/xsd/charset.rb', line 91

def Charset.encoding=(encoding)
  warn("xsd charset is set to #{encoding}") if $DEBUG
  @internal_encoding = encoding
end

.encoding_conv(str, enc_from, enc_to) ⇒ Object


108
109
110
111
112
113
114
115
116
117
# File 'lib/xsd/charset.rb', line 108

def Charset.encoding_conv(str, enc_from, enc_to)
  if enc_from == enc_to or enc_from == 'NONE' or enc_to == 'NONE'
    str
  elsif converter = EncodingConvertMap[[enc_from, enc_to]]
    converter.call(str)
  else
    raise CharsetConversionError.new(
	"Converter not found: #{enc_from} -> #{enc_to}")
  end
end

.encoding_from_xml(str, charset) ⇒ Object


104
105
106
# File 'lib/xsd/charset.rb', line 104

def Charset.encoding_from_xml(str, charset)
  encoding_conv(str, charset_str(charset), @internal_encoding)
end

.encoding_to_xml(str, charset) ⇒ Object


100
101
102
# File 'lib/xsd/charset.rb', line 100

def Charset.encoding_to_xml(str, charset)
  encoding_conv(str, @internal_encoding, charset_str(charset))
end

.initObject


27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
# File 'lib/xsd/charset.rb', line 27

def Charset.init
  EncodingConvertMap[['UTF8', 'X_ISO_8859_1']] =
    Proc.new { |str| str.unpack('U*').pack('C*') }
  EncodingConvertMap[['X_ISO_8859_1', 'UTF8']] =
    Proc.new { |str| str.unpack('C*').pack('U*') }
  begin
    require 'xsd/iconvcharset'
    @internal_encoding = 'UTF8'
    sjtag = (/(mswin|bccwin|mingw|cygwin|emx)/ =~ RUBY_PLATFORM) ? 'cp932' :
      'shift_jis'
    EncodingConvertMap[['UTF8', 'EUC' ]] =
      Proc.new { |str| IconvCharset.safe_iconv("euc-jp", "utf-8", str) }
    EncodingConvertMap[['EUC' , 'UTF8']] =
      Proc.new { |str| IconvCharset.safe_iconv("utf-8", "euc-jp", str) }
    EncodingConvertMap[['EUC' , 'SJIS']] =
      Proc.new { |str| IconvCharset.safe_iconv(sjtag, "euc-jp", str) }
    EncodingConvertMap[['UTF8', 'SJIS']] =
      Proc.new { |str| IconvCharset.safe_iconv(sjtag, "utf-8", str) }
    EncodingConvertMap[['SJIS', 'UTF8']] =
      Proc.new { |str| IconvCharset.safe_iconv("utf-8", sjtag, str) }
    EncodingConvertMap[['SJIS', 'EUC' ]] =
      Proc.new { |str| IconvCharset.safe_iconv("euc-jp", sjtag, str) }
  rescue LoadError
    begin
     	require 'nkf'
	EncodingConvertMap[['EUC' , 'SJIS']] =
        Proc.new { |str| NKF.nkf('-sXm0', str) }
	EncodingConvertMap[['SJIS', 'EUC' ]] =
        Proc.new { |str| NKF.nkf('-eXm0', str) }
    rescue LoadError
    end

    begin
	require 'uconv'
	@internal_encoding = 'UTF8'
	EncodingConvertMap[['UTF8', 'EUC' ]] = Uconv.method(:u8toeuc)
	EncodingConvertMap[['UTF8', 'SJIS']] = Uconv.method(:u8tosjis)
	EncodingConvertMap[['EUC' , 'UTF8']] = Uconv.method(:euctou8)
	EncodingConvertMap[['SJIS', 'UTF8']] = Uconv.method(:sjistou8)
    rescue LoadError
    end
  end
end

.is_ces(str, code = @internal_encoding) ⇒ Object


173
174
175
176
177
178
179
180
181
182
183
184
185
186
# File 'lib/xsd/charset.rb', line 173

def Charset.is_ces(str, code = @internal_encoding)
  case code
  when 'NONE'
    is_us_ascii(str)
  when 'UTF8'
    is_utf8(str)
  when 'EUC'
    is_euc(str)
  when 'SJIS'
    is_sjis(str)
  else
    raise UnknownCharsetError.new("Unknown charset: #{code}")
  end
end

.is_euc(str) ⇒ Object


165
166
167
# File 'lib/xsd/charset.rb', line 165

def Charset.is_euc(str)
  EUCRegexp =~ str
end

.is_sjis(str) ⇒ Object


169
170
171
# File 'lib/xsd/charset.rb', line 169

def Charset.is_sjis(str)
  SJISRegexp =~ str
end

.is_us_ascii(str) ⇒ Object


157
158
159
# File 'lib/xsd/charset.rb', line 157

def Charset.is_us_ascii(str)
  USASCIIRegexp =~ str
end

.is_utf8(str) ⇒ Object


161
162
163
# File 'lib/xsd/charset.rb', line 161

def Charset.is_utf8(str)
  UTF8Regexp =~ str
end

.xml_encoding_labelObject


96
97
98
# File 'lib/xsd/charset.rb', line 96

def Charset.xml_encoding_label
  charset_label(@internal_encoding)
end