# -*- coding: utf-8 -*- =begin mo.rb - A simple class for operating GNU MO file. Copyright (C) 2012 Kouhei Sutou Copyright (C) 2003-2009 Masao Mutoh Copyright (C) 2002 Masahiro Sakai, Masao Mutoh Copyright (C) 2001 Masahiro Sakai Masahiro Sakai Masao Mutoh You can redistribute this file and/or modify it under the same term of Ruby. License of Ruby is included with Ruby distribution in the file "README". =end require 'stringio' module GetText class MO < Hash class InvalidFormat < RuntimeError; end; attr_reader :filename Header = Struct.new(:magic, :revision, :nstrings, :orig_table_offset, :translated_table_offset, :hash_table_size, :hash_table_offset) # The following are only used in .mo files # with minor revision >= 1. class HeaderRev1 < Header attr_accessor :n_sysdep_segments, :sysdep_segments_offset, :n_sysdep_strings, :orig_sysdep_tab_offset, :trans_sysdep_tab_offset end MAGIC_BIG_ENDIAN = "\x95\x04\x12\xde".force_encoding("ASCII-8BIT") MAGIC_LITTLE_ENDIAN = "\xde\x12\x04\x95".force_encoding("ASCII-8BIT") def self.open(arg = nil, output_charset = nil) result = self.new(output_charset) result.load(arg) end def initialize(output_charset = nil) @filename = nil @last_modified = nil @little_endian = true @output_charset = output_charset @plural_proc = nil super() end def store(msgid, msgstr, options) string = generate_original_string(msgid, options) self[string] = msgstr end def update! if FileTest.exist?(@filename) st = File.stat(@filename) load(@filename) unless (@last_modified == [st.ctime, st.mtime]) else warn "#{@filename} was lost." if $DEBUG clear end self end def load(arg) if arg.kind_of? String begin st = File.stat(arg) @last_modified = [st.ctime, st.mtime] rescue Exception end load_from_file(arg) else load_from_stream(arg) end @filename = arg self end def load_from_stream(io) magic = io.read(4) case magic when MAGIC_BIG_ENDIAN @little_endian = false when MAGIC_LITTLE_ENDIAN @little_endian = true else raise InvalidFormat.new(sprintf("Unknown signature %s", magic.dump)) end endian_type6 = @little_endian ? 'V6' : 'N6' endian_type_astr = @little_endian ? 'V*' : 'N*' header = HeaderRev1.new(magic, *(io.read(4 * 6).unpack(endian_type6))) if header.revision == 1 # FIXME: It doesn't support sysdep correctly. header.n_sysdep_segments = io.read(4).unpack(endian_type6) header.sysdep_segments_offset = io.read(4).unpack(endian_type6) header.n_sysdep_strings = io.read(4).unpack(endian_type6) header.orig_sysdep_tab_offset = io.read(4).unpack(endian_type6) header.trans_sysdep_tab_offset = io.read(4).unpack(endian_type6) elsif header.revision > 1 raise InvalidFormat.new(sprintf("file format revision %d isn't supported", header.revision)) end io.pos = header.orig_table_offset orig_table_data = io.read((4 * 2) * header.nstrings).unpack(endian_type_astr) io.pos = header.translated_table_offset trans_table_data = io.read((4 * 2) * header.nstrings).unpack(endian_type_astr) msgids = Array.new(header.nstrings) for i in 0...header.nstrings io.pos = orig_table_data[i * 2 + 1] msgids[i] = io.read(orig_table_data[i * 2 + 0]) end clear for i in 0...header.nstrings io.pos = trans_table_data[i * 2 + 1] msgstr = io.read(trans_table_data[i * 2 + 0]) msgid = msgids[i] if msgid.nil? || msgid.empty? if msgstr @charset = nil @nplurals = nil @plural = nil msgstr.each_line{|line| if /^Content-Type:/i =~ line and /charset=((?:\w|-)+)/i =~ line @charset = $1 elsif /^Plural-Forms:\s*nplurals\s*\=\s*(\d*);\s*plural\s*\=\s*([^;]*)\n?/ =~ line @nplurals = $1 @plural = $2 end break if @charset and @nplurals } @nplurals = "1" unless @nplurals @plural = "0" unless @plural end else unless msgstr.nil? msgstr = convert_encoding(msgstr, msgid) end end self[convert_encoding(msgid, msgid)] = msgstr.freeze end self end def prime?(number) ('1' * number) !~ /^1?$|^(11+?)\1+$/ end begin require 'prime' def next_prime(seed) Prime.instance.find{|x| x > seed } end rescue LoadError def next_prime(seed) require 'mathn' prime = Prime.new while current = prime.succ return current if current > seed end end end HASHWORDBITS = 32 # From gettext-0.12.1/gettext-runtime/intl/hash-string.h # Defines the so called `hashpjw' function by P.J. Weinberger # [see Aho/Sethi/Ullman, COMPILERS: Principles, Techniques and Tools, # 1986, 1987 Bell Telephone Laboratories, Inc.] def hash_string(str) hval = 0 str.each_byte do |b| break if b == '\0' hval <<= 4 hval += b.to_i g = hval & (0xf << (HASHWORDBITS - 4)) if (g != 0) hval ^= g >> (HASHWORDBITS - 8) hval ^= g end end hval end #Save data as little endian format. def save_to_stream(io) # remove untranslated message translated_messages = reject do |msgid, msgstr| msgstr.nil? end size = translated_messages.size header_size = 4 * 7 table_size = 4 * 2 * size hash_table_size = next_prime((size * 4) / 3) hash_table_size = 3 if hash_table_size <= 2 header = Header.new( MAGIC_LITTLE_ENDIAN, # magic 0, # revision size, # nstrings header_size, # orig_table_offset header_size + table_size, # translated_table_offset hash_table_size, # hash_table_size header_size + table_size * 2 # hash_table_offset ) io.write(header.to_a.pack('a4V*')) ary = translated_messages.to_a ary.sort!{|a, b| a[0] <=> b[0]} # sort by original string pos = header.hash_table_size * 4 + header.hash_table_offset orig_table_data = Array.new() ary.each{|item, _| orig_table_data.push(item.bytesize) orig_table_data.push(pos) pos += item.bytesize + 1 # +1 is } io.write(orig_table_data.pack('V*')) trans_table_data = Array.new() ary.each{|_, item| trans_table_data.push(item.bytesize) trans_table_data.push(pos) pos += item.bytesize + 1 # +1 is } io.write(trans_table_data.pack('V*')) hash_tab = Array.new(hash_table_size) j = 0 ary[0...size].each {|key, _| hash_val = hash_string(key) idx = hash_val % hash_table_size if hash_tab[idx] != nil incr = 1 + (hash_val % (hash_table_size - 2)) begin if (idx >= hash_table_size - incr) idx -= hash_table_size - incr else idx += incr end end until (hash_tab[idx] == nil) end hash_tab[idx] = j + 1 j += 1 } hash_tab.collect!{|i| i ? i : 0} io.write(hash_tab.pack('V*')) ary.each{|item, _| io.write(item); io.write("\0") } ary.each{|_, item| io.write(item); io.write("\0") } self end def load_from_file(filename) @filename = filename begin File.open(filename, 'rb'){|f| load_from_stream(f)} rescue => e e.set_backtrace("File: #{@filename}") raise e end end def save_to_file(filename) File.open(filename, 'wb'){|f| save_to_stream(f)} end def set_comment(msgid_or_sym, comment) #Do nothing end def plural_as_proc unless @plural_proc @plural_proc = Proc.new{|n| eval(@plural)} begin @plural_proc.call(1) rescue @plural_proc = Proc.new{|n| 0} end end @plural_proc end attr_accessor :little_endian, :path, :last_modified attr_reader :charset, :nplurals, :plural private def convert_encoding(string, original_string) return string if @output_charset.nil? or @charset.nil? if Encoding.find(@output_charset) == Encoding.find(@charset) string.force_encoding(@output_charset) return string end begin string.encode(@output_charset, @charset) rescue EncodingError if $DEBUG warn "@charset = ", @charset warn "@output_charset = ", @output_charset warn "msgid = ", original_string warn "msgstr = ", string end string end end def generate_original_string(msgid, options) string = "" msgctxt = options.delete(:msgctxt) msgid_plural = options.delete(:msgid_plural) string << msgctxt << "\004" unless msgctxt.nil? string << msgid string << "\000" << msgid_plural unless msgid_plural.nil? string end end end # Test if $0 == __FILE__ if (ARGV.include? "-h") or (ARGV.include? "--help") STDERR.puts("mo.rb [filename.mo ...]") exit end ARGV.each{ |item| mo = GetText::MO.open(item) puts "------------------------------------------------------------------" puts "charset = \"#{mo.charset}\"" puts "nplurals = \"#{mo.nplurals}\"" puts "plural = \"#{mo.plural}\"" puts "------------------------------------------------------------------" mo.each do |key, value| puts "original message = #{key.inspect}" puts "translated message = #{value.inspect}" puts "--------------------------------------------------------------------" end } end