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,
}
USASCIIRegexp =

XML 1.0 restricted.

Regexp.new("\\A#{us_ascii}*\\z", nil, "NONE")
EUCRegexp =
Regexp.new("\\A#{character_euc}*\\z", nil, "NONE")
SJISRegexp =
Regexp.new("\\A#{character_sjis}*\\z", nil, "NONE")
UTF8Regexp =
Regexp.new("\\A#{character_utf8}*\\z", nil, "NONE")

Class Method Summary collapse

Class Method Details

.charset_label(encoding) ⇒ Object


116
117
118
# File 'lib/xsd/charset.rb', line 116

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

.charset_str(label) ⇒ Object


120
121
122
123
124
125
126
# File 'lib/xsd/charset.rb', line 120

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

.encodingObject

handlers


84
85
86
# File 'lib/xsd/charset.rb', line 84

def Charset.encoding
  @internal_encoding
end

.encoding=(encoding) ⇒ Object


88
89
90
91
# File 'lib/xsd/charset.rb', line 88

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


105
106
107
108
109
110
111
112
113
114
# File 'lib/xsd/charset.rb', line 105

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


101
102
103
# File 'lib/xsd/charset.rb', line 101

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

.encoding_to_xml(str, charset) ⇒ Object


97
98
99
# File 'lib/xsd/charset.rb', line 97

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

.initObject


26
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
# File 'lib/xsd/charset.rb', line 26

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


170
171
172
173
174
175
176
177
178
179
180
181
182
183
# File 'lib/xsd/charset.rb', line 170

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


162
163
164
# File 'lib/xsd/charset.rb', line 162

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

.is_sjis(str) ⇒ Object


166
167
168
# File 'lib/xsd/charset.rb', line 166

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

.is_us_ascii(str) ⇒ Object


154
155
156
# File 'lib/xsd/charset.rb', line 154

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

.is_utf8(str) ⇒ Object


158
159
160
# File 'lib/xsd/charset.rb', line 158

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

.xml_encoding_labelObject


93
94
95
# File 'lib/xsd/charset.rb', line 93

def Charset.xml_encoding_label
  charset_label(@internal_encoding)
end