#! ruby -Ku
  # mkrexml.rb version 0.8  (coding: utf-8)  2010/10/27
  # Copyright (C) T. Yoshiizumi, 2009-2010 All rights reserved.
  # $ ruby mkrexml.rb input.xml (this will create input_xml.rb )
  # $ ruby input_xml.rb > output.xml (which is about the same as input.xml)
  # the encoding of output.xml shall be utf-8
require "rexml/document"

Mkrexml_DS = '_LI842zWB_'
Mkrexml_DE = '_BFD4WGnH_'

# dump with marking multi-byte characters
def k_dump(s)
  s = s.gsub(/[^\x00-\x7f]+/) {|mp|
    mp = "#{Mkrexml_DS}#{mp}#{Mkrexml_DE}"
  }
  return s.dump
end

# recover a string dumped by k_dump
def k_undump(s)
  s = s.gsub(/#{Mkrexml_DS}(.+?)#{Mkrexml_DE}/) {|mp|
    w = '"' + $1 + '"'
    mp = eval(w)
  }
  return s
end


# element name -> ruby variable name
def var_name(str)
  res = str.gsub(/[A-Za-z]+/) {|part|
    part = part.gsub(/([A-Z])([a-z])/) {|mp|
      mp = "_" + $1.tr("A-Z", "a-z") + $2
    }
    part = part.gsub(/([A-Z]+)/) {|mp|
      mp = "_" + mp.tr("A-Z", "a-z")
    }
    part = part.gsub(/__+/, "_")
    part = part.sub(/^_/, "")
  }
  res = "@" + res.gsub(/[^0-9A-Za-z_]/, "_")
end


# make a script on each element (recurse)
def seek_element(element, parent)
  if parent.class == String  # the element is root(parent is not an element)
    parent_name = parent
  else  # if parent.class == REXML::Element
    parent_name = var_name(parent.expanded_name)
  end
  element_name = var_name(element.expanded_name)
  hs_str = ""
  if element.has_attributes?
    element.attributes.each {|key,val|
      hs_str += sprintf("\t%s=>%s,\n", k_dump(key), k_dump(val))
    }
    hs_str = hs_str.sub(/\A\t+/, "{")
    hs_str = hs_str.sub(/,\n\Z/, "}")
  end
  text_str = ""
  if element.has_text?
    text_str = element.text.to_s
  end
      # make a script string
  scr = ""
  if hs_str == ""
    scr = "#{element_name} = #{parent_name}.add_element(#{k_dump(element.expanded_name)})"
  else
    scr = "#{element_name} = #{parent_name}." +
        "add_element(#{k_dump(element.expanded_name)}, #{hs_str})"
  end
  if text_str != ""
    if text_str =~ /^\s+$/
      scr += ".add_text(#{k_dump(text_str)})"
    else
      scr += "\n#{element_name}.add_text(#{k_dump(text_str)})"
    end
  end
  if scr != ""
    scr = scr + "\n\n"
  end
      # recurse
  if element.elements.size > 0
    element.elements.each {|c_element|
      scr = scr + seek_element(c_element, element)
    }
  end
  return scr
end


## main
ARGV.each {|infile|  # resource xml file
  outrb = infile.gsub(/\./, "_") + ".rb"
  doc_name = "doc"  # name of REXML::Document in the outrb

  src = File.open(infile) {|fp|  fp.read}
  doc = REXML::Document.new(src)
  doc.xml_decl.encoding = "UTF-8"

  rb_str = <<EOS
#! ruby -Ku
  # coding: utf-8
  # #{File.basename(infile)} -> #{File.basename(outrb)} at #{Time.now.strftime("%Y/%m/%d %H:%M")}
require \"rexml/document\"
EOS

  before_str = ""
  after_str = ""
  element_str = ""
  other_str = ""
  doc.each {|e|
    unless e.class == REXML::Element
      if element_str == ""
        before_str += e.to_s
      elsif e.class == REXML::Text
        after_str += e.to_s
      else
        other_str += sprintf("- %s: <%s>, %s\n", infile, e.class.to_s, e.to_s)
      end
    else  # if e.class == REXML::Element
      element_str += seek_element(e, doc_name)
    end
  }
  rb_str += "#{doc_name} = REXML::Document.new(#{k_dump(before_str)})\n\n"
  rb_str += element_str
  rb_str += "#{doc_name}.add_text(#{k_dump(after_str)})\n"
  rb_str += "print #{doc_name}\n"
  rb_str = k_undump(rb_str)
  File.open(outrb, "w") {|ff|  ff.write rb_str}
  if other_str != ""
    STDERR.puts "** missing the infomation below\n#{other_str}"
  end
}