#!/usr/bin/env ruby
# simpleurl.rb
#
# TAKAHASHI 'Maki' Masayoshi
# It can use as public domain software.
#
# これは、Tomoyuki Kosimizuさんのuri.rb
# http://www.ruby-lang.org/en/raa-list.rhtml?name=uri.rb
# http://www02.so-net.ne.jp/%7Egreentea/ruby/uri/
# を元に、大幅に短縮化・機能制限をかけたものです。
#
# original id:
#    uri.rb,v 4.22 2000/08/19 04:12:04 rgt Exp

=begin

== URI class
 --- URI.create(url_string)
       create URI object.

 --- URI.escape(str, unsafe=nil)
 --- URI.unescape(str)

 --- URI#==(obj)
 --- URI#to_s()

 --- URI#scheme
 --- URI#host
 --- URI#host=(str)
 --- URI#port
 --- URI#port=(str)
 --- URI#path
 --- URI#path=(str)
 --- URI#query
 --- URI#query=(str)
 --- URI#fragment
 --- URI#fragment=(str)
 --- URI#userinfo
 --- URI#userinfo=(str)
 
=end

=begin
  To escape query string, use URI#escape. (Thanx to NaHi, Wakou)
=end

class URIError < StandardError; end
class URI

  ##pre absolute_uri and ((relative_uri.nil? and
  ##                       absolute_uri.is_a? String) or
  ##                      (relative_uri.is_a? String and 
  ##                       (absolute_uri.is_a? String or
  ##                        absolute_uri.is_a? URI)))
  ##post const absolute_uri, const relative_uri
  ##return aURI
  ##raise URIError if absolute_uri and/or relative_uri are a bad URI

  def URI.create(absolute_uri)
    if absolute_uri.is_a? URI
      absolute_uri
    else
      URI.new(absolute_uri)
    end
  end

=begin
  escape str.
  default unsafe is ^uric. see URI::Escape.pm in perl5.
  if you want to escape str for CGI, use CGI::escape.
=end
  ##pre str.is_a? String and (unsafe.nil? or unsafe.is_a? String)
  def URI.escape(str, unsafe=nil)
    if not unsafe
      str.gsub(/[^;\/?:@&=+$,0-9A-Za-z\-_.!~*'()\[\]]/) do |match|  #'
	sprintf('%%%02X', match.unpack('C')[0])
      end
    else
      str.gsub(/[#{unsafe}]/) do |match|
	sprintf('%%%02X', match.unpack('C')[0])
      end
    end
  end

  def URI.unescape(str)
    result = str.dup
    while /%[0-9A-Fa-f][0-9A-Fa-f]/ =~ result
      char = $&[1,2].hex.chr
      result.sub!(/%[0-9A-Fa-f][0-9A-Fa-f]/, char)
    end
    result
  end

  attr_reader :scheme
  attr_accessor :host, :path, :query, :fragment, :userinfo, :port

  def initialize(uri)
    if %r|\A(([^:/?#]+):)?(//([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?\z| !~ uri
	#     12            3  4          5       6  7        8 9
	# scheme = $2, authority = $4, path = $5, query = $7, fragment = $9
      raise URIError('bad URI: ' + uri)
    end
    @scheme, @authority, @path, @query, @fragment =
     $2,      $4,         $5,    $7,     $9

    if @authority.nil?
      raise URIError, "undefined authority in '#{uri}'"
    end
    @userinfo, @host, @port = parse_server(@authority)
    if not @path or @path == ''
      @path = '/'
    else
      @path = resolve_path(@path)
    end
  end

  # escape all chars except escaped
  ##private
  def escape_all(str)
    return nil if str.nil?
    tmp = nil
    buff = ''
    arry = str.split(//)
    i = 0
    while i < arry.size
      if arry[i] == '%'
	(tmp = arry[0, 3].join('')).upcase!
	if /%[0-9A-Z][0-9A-Z]/ =~ tmp
	  buff << tmp
	  i += 3
	  next
	else
	  buff << sprintf("%%%02X", ?%)
	end
      else
	buff << sprintf("%%%02X", arry[i][0])
      end
      i += 1
    end
    buff
  end
  private :escape_all

  # userinfo@host:port
  ##private
  def parse_server(authority)
    userinfo, host, port =
      authority.scan(/\A(?:([^@]*)@)?([^:]*)(?::(.*))?\z/)[0]
    [userinfo, host, port]
  end
  private :parse_server

  # resolve absolute path
  # see RFC-2396, C.2. Abnormal Examples
  ##private
  def resolve_path(path)
    if path == '/./' or path == '/.' or path == '/../' or path == '/..'
      return path.dup
    end
    result = remove_current_dir path
    remove_parent_dir result
  end
  private :resolve_path

  ##private
  def remove_current_dir(path)
    result = path.dup
    if %r|/\.\.?\z| =~ path
      result << '/'
    end
    result.gsub!(%r|/(\./)+|, '/')
    if path[0] == ?/ and path[1] == ?. and path[2] == ?/
      result = '/.' << result
    end
    result
  end
  private :remove_current_dir

  ##private
  def remove_parent_dir(apath)
    result = []
    apath.split(%r|/|, -1).each do |elt|
      if elt == '..'
	if result.size == 1 or result[-1] == dOT_DOT
	  result.push elt
	else
	  result.pop
	end
      else
	result.push elt
      end
    end
    result.join('/')
  end
  private :remove_parent_dir

  def ==(obj)
    m = y = nil  # mine, yours
    if obj.is_a? String
      begin
	obj = URI.create(obj)
      rescue URIError
	return false
      end
    end
    if not super(obj) or
	(@userinfo != obj.userinfo) or
	(escape_all(@host) != escape_all(obj.userinfo)) or
	(escape_all(@query) != escape_all(obj.query)) or
	(@port != obj.port) or
	(escape_all(@fragment) != escape_all(obj.fragment)) or
	(not_equal_path? obj)
      return false
    end
    true
  end

  ##private
  def not_equal_path?(other)
    return false if @path == other.path

    arry_a = @path.split('/', -1)
    arry_b = other.path.split('/', -1)
    return true if arry_a.size != arry_b.size

    arry_a.size.times do |i|
      if arry_a[i] != arry_b[i] and
	  escape_all(arry_a[i]) != escape_all(arry_b[i])
	return true
      end
    end
    false
  end
  private :not_equal_path?

  def to_s
    tmp = nil
    buff = @scheme + '://'
    if @userinfo
      buff << @userinfo << '@'
    end
    buff << @host
    if @port
      buff << ':' << @port
    end
    buff << @path
    if @query
      buff << '?' << @query
    end
    if @fragment
      buff << '#' << @fragment
    end
    buff
  end
end