1  # -*- mode: ruby; ruby-indent-level: 2; indent-tabs-mode: t; tab-width: 2 -*- vim: sw=2 ts=2
   2  module REXML
   3    module Encoding
   4      @encoding_methods = {}
   5      def self.register(enc, &block)
   6        @encoding_methods[enc] = block
   7      end
   8      def self.apply(obj, enc)
   9        @encoding_methods[enc][obj]
  10      end
  11      def self.encoding_method(enc)
  12        @encoding_methods[enc]
  13      end
  14 
  15      # Native, default format is UTF-8, so it is declared here rather than in
  16      # an encodings/ definition.
  17      UTF_8 = 'UTF-8'
  18      UTF_16 = 'UTF-16'
  19      UNILE = 'UNILE'
  20 
  21      # ID ---> Encoding name
  22      attr_reader :encoding
  23      def encoding=( enc )
  24        old_verbosity = $VERBOSE
  25        begin
  26          $VERBOSE = false
  27          enc = enc.nil? ? nil : enc.upcase
  28          return false if defined? @encoding and enc == @encoding
  29          if enc and enc != UTF_8
  30            @encoding = enc
  31            raise ArgumentError, "Bad encoding name #@encoding" unless @encoding =~ /^[\w-]+$/
  32            @encoding.untaint 
  33            begin
  34              require 'rexml/encodings/ICONV.rb'
  35              Encoding.apply(self, "ICONV")
  36            rescue LoadError, Exception
  37              begin
  38                enc_file = File.join( "rexml", "encodings", "#@encoding.rb" )
  39                require enc_file
  40                Encoding.apply(self, @encoding)
  41              rescue LoadError => err
  42                puts err.message
  43                raise ArgumentError, "No decoder found for encoding #@encoding.  Please install iconv."
  44              end
  45            end
  46          else
  47            @encoding = UTF_8
  48            require 'rexml/encodings/UTF-8.rb'
  49            Encoding.apply(self, @encoding)
  50          end
  51        ensure
  52          $VERBOSE = old_verbosity
  53        end
  54        true
  55      end
  56 
  57      def check_encoding str
  58        # We have to recognize UTF-16, LSB UTF-16, and UTF-8
  59        if str[0] == 0xfe && str[1] == 0xff
  60          str[0,2] = ""
  61          return UTF_16
  62        elsif str[0] == 0xff && str[1] == 0xfe
  63          str[0,2] = ""
  64          return UNILE
  65        end
  66        str =~ /^\s*<\?xml\s+version\s*=\s*(['"]).*?\1\s+encoding\s*=\s*(["'])(.*?)\2/um
  67        return $3.upcase if $3
  68        return UTF_8
  69      end
  70    end
  71  end