# -*- mode: ruby; coding: utf-8 -*- # # po_parser.ry - ruby version of msgfmt # # Copyright (C) 2002-2008 Masao Mutoh # Copyright (C) 2012-2023 Sutou Kouhei # Copyright (C) 2012-2013 Haruka Yoshihara # # You may redistribute it and/or modify it under the same # license terms as Ruby or LGPL. class GetText::POParser token COMMENT MSGID MSGCTXT MSGID_PLURAL MSGSTR STRING PLURAL_NUM rule msgfmt : /* empty */ | msgfmt comment | msgfmt msgctxt | msgfmt message ; msgctxt : MSGCTXT string_list { @msgctxt = val[1] } ; message : single_message | plural_message ; single_message : MSGID string_list MSGSTR string_list { msgid = val[1] msgstr = val[3] use_message_p = true if @fuzzy and not msgid.empty? use_message_p = (not ignore_fuzzy?) if report_warning? if ignore_fuzzy? $stderr.print _("Warning: fuzzy message was ignored.\n") else $stderr.print _("Warning: fuzzy message was used.\n") end $stderr.print " #{@po_file}: msgid '#{msgid}'\n" end end @fuzzy = false if use_message_p on_message(msgid, msgstr) else clear end result = "" } plural_message : MSGID string_list MSGID_PLURAL string_list msgstr_plural { if @fuzzy and ignore_fuzzy? if val[1] != "" if report_warning? $stderr.print _("Warning: fuzzy message was ignored.\n") $stderr.print "msgid = '#{val[1]}\n" end else on_message("", unescape(val[3])) end @fuzzy = false else @msgid_plural = unescape(val[3]) on_message(unescape(val[1]), unescape(val[4])) end result = "" } ; msgstr_plural : msgstr_plural msgstr_plural_line { if val[0].size > 0 result = val[0] + "\000" + val[1] else result = "" end } | msgstr_plural_line ; msgstr_plural_line : MSGSTR PLURAL_NUM string_list { result = val[2] } ; comment : COMMENT { on_comment(val[0]) } #| COMMENT #; string_list : string_list STRING { result = val.delete_if{|item| item == ""}.join } | STRING { result = val[0] } ; end ---- header require "gettext/po" ---- inner if GetText.respond_to?(:bindtextdomain) include GetText GetText.bindtextdomain("gettext") else def _(message_id) message_id end private :_ end attr_writer :ignore_fuzzy, :report_warning def initialize @ignore_fuzzy = true @report_warning = true end def ignore_fuzzy? @ignore_fuzzy end def report_warning? @report_warning end def unescape(string) string.gsub(/\\(.)/) do escaped_character = $1 case escaped_character when "t" "\t" when "r" "\r" when "n" "\n" else escaped_character end end end private :unescape def parse(str, data) clear @data = data str = str.strip @q = [] until str.empty? do case str when /\A\s+/ str = $' when /\Amsgctxt/ @q.push [:MSGCTXT, $&] str = $' when /\Amsgid_plural/ @q.push [:MSGID_PLURAL, $&] str = $' when /\Amsgid/ @q.push [:MSGID, $&] str = $' when /\Amsgstr/ @q.push [:MSGSTR, $&] str = $' when /\A\[(\d+)\]/ @q.push [:PLURAL_NUM, $1] str = $' when /\A\#~(.*)/ if report_warning? $stderr.print _("Warning: obsolete msgid exists.\n") $stderr.print " #{$&}\n" end @q.push [:COMMENT, $&] str = $' when /\A\#(.*)/ @q.push [:COMMENT, $&] str = $' when /\A\"(.*)\"/ @q.push [:STRING, unescape($1)] str = $' else #c = str[0,1] #@q.push [:STRING, c] str = str[1..-1] end end @q.push [false, "$end"] if $DEBUG @q.each do |a,b| puts "[#{a}, #{b}]" end end @yydebug = true if $DEBUG do_parse if @comments.size > 0 @data.set_comment(:last, @comments.join("\n")) end @data end def next_token @q.shift end def on_message(msgid, msgstr) msgstr = nil if msgstr.empty? if @data.instance_of?(PO) type = detect_entry_type entry = POEntry.new(type) entry.translator_comment = format_comment(@translator_comments) entry.extracted_comment = format_comment(@extracted_comments) entry.flags = @flags entry.previous = format_comment(@previous) entry.references = @references entry.msgctxt = @msgctxt entry.msgid = msgid entry.msgid_plural = @msgid_plural entry.msgstr = msgstr @data[@msgctxt, msgid] = entry else options = {} options[:msgctxt] = @msgctxt options[:msgid_plural] = @msgid_plural @data.store(msgid, msgstr, options) @data.set_comment(msgid, format_comment(@comments)) end clear end def format_comment(comments) return "" if comments.empty? comment = comments.join("\n") comment << "\n" if comments.last.empty? comment end def on_comment(comment) @fuzzy = true if (/fuzzy/ =~ comment) if @data.instance_of?(PO) if comment == "#" @translator_comments << "" elsif /\A(#.)\s*(.*)\z/ =~ comment mark = $1 content = $2 case mark when POFormat::TRANSLATOR_COMMENT_MARK @translator_comments << content when POFormat::EXTRACTED_COMMENT_MARK @extracted_comments << content when POFormat::REFERENCE_COMMENT_MARK @references.concat(parse_references_line(content)) when POFormat::FLAG_MARK @flags.concat(parse_flags_line(content)) when POFormat::PREVIOUS_COMMENT_MARK @previous << content else @comments << comment end end else @comments << comment end end def parse_file(po_file, data) args = [ po_file ] # In Ruby 1.9, we must detect proper encoding of a PO file. if String.instance_methods.include?(:encode) encoding = detect_file_encoding(po_file) args << "r:#{encoding}" end @po_file = po_file parse(File.open(*args) {|io| io.read }, data) end private def detect_file_encoding(po_file) open(po_file, :encoding => "ASCII-8BIT") do |input| in_header = false input.each_line do |line| case line.chomp when /\Amsgid\s+"(.*)"\z/ id = $1 break unless id.empty? in_header = true when /\A"Content-Type:.*\scharset=(.*)\\n"\z/ charset = $1 next unless in_header break if template_charset?(charset) return Encoding.find(charset) end end end Encoding.default_external end def template_charset?(charset) charset == "CHARSET" end def detect_entry_type if @msgctxt.nil? if @msgid_plural.nil? :normal else :plural end else if @msgid_plural.nil? :msgctxt else :msgctxt_plural end end end def parse_references_line(line) line.split(/\s+/) end def parse_flags_line(line) line.split(/\s+/) end def clear @translator_comments = [] @extracted_comments = [] @references = [] @flags = [] @previous = [] @references = [] @comments = [] @msgctxt = nil @msgid_plural = nil end ---- footer