module RDoc::Encoding

This class is a wrapper around File IO and Encoding that helps RDoc load files and convert them to the correct encoding.

Public Class Methods

change_encoding(text, encoding) click to toggle source

Changes encoding based on encoding without converting and returns new string

# File lib/rdoc/encoding.rb, line 112
def self.change_encoding text, encoding
  if text.kind_of? RDoc::Comment
    text.encode! encoding
  else
    String.new text, encoding: encoding
  end
end
detect_encoding(string) click to toggle source

Detects the encoding of string based on the magic comment

# File lib/rdoc/encoding.rb, line 92
def self.detect_encoding string
  result = HEADER_REGEXP.match string
  name = result && result[:name]

  name ? Encoding.find(name) : nil
end
read_file(filename, encoding, force_transcode = false) click to toggle source

Reads the contents of filename and handles any encoding directives in the file.

The content will be converted to the encoding. If the file cannot be converted a warning will be printed and nil will be returned.

If force_transcode is true the document will be transcoded and any unknown character in the target encoding will be replaced with ‘?’

# File lib/rdoc/encoding.rb, line 32
def self.read_file filename, encoding, force_transcode = false
  content = File.open filename, "rb" do |f| f.read end
  content.gsub!("\r\n", "\n") if RUBY_PLATFORM =~ /mswin|mingw/

  utf8 = content.sub!(/\A\xef\xbb\xbf/, '')

  enc = RDoc::Encoding.detect_encoding content
  content = RDoc::Encoding.change_encoding content, enc if enc

  begin
    encoding ||= Encoding.default_external
    orig_encoding = content.encoding

    if not orig_encoding.ascii_compatible? then
      content = content.encode encoding
    elsif utf8 then
      content = RDoc::Encoding.change_encoding content, Encoding::UTF_8
      content = content.encode encoding
    else
      # assume the content is in our output encoding
      content = RDoc::Encoding.change_encoding content, encoding
    end

    unless content.valid_encoding? then
      # revert and try to transcode
      content = RDoc::Encoding.change_encoding content, orig_encoding
      content = content.encode encoding
    end

    unless content.valid_encoding? then
      warn "unable to convert #{filename} to #{encoding}, skipping"
      content = nil
    end
  rescue Encoding::InvalidByteSequenceError,
         Encoding::UndefinedConversionError => e
    if force_transcode then
      content = RDoc::Encoding.change_encoding content, orig_encoding
      content = content.encode(encoding,
                               :invalid => :replace,
                               :undef => :replace,
                               :replace => '?')
      return content
    else
      warn "unable to convert #{e.message} for #{filename}, skipping"
      return nil
    end
  end

  content
rescue ArgumentError => e
  raise unless e.message =~ /unknown encoding name - (.*)/
  warn "unknown encoding name \"#{$1}\" for #{filename}, skipping"
  nil
rescue Errno::EISDIR, Errno::ENOENT
  nil
end
remove_magic_comment(string) click to toggle source

Removes magic comments and shebang

# File lib/rdoc/encoding.rb, line 102
def self.remove_magic_comment string
  string.sub HEADER_REGEXP do |s|
    s.gsub(/[^\n]/, '')
  end
end