added uri library (uri-0.9.4)

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@1979 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

lib/README

@@ -61,4 +61,10 @@ thread.rb	thread support
 thwait.rb	thread syncronization class
 timeout.rb	provides timeout
 tracer.rb	execution tracer
+uri.rb		URI support
+uri/ftp.rb	ftp scheme support
+uri/http.rb	http scheme support
+uri/https.rb	https scheme support
+uri/ldap.rb	ldap scheme support
+uri/mailto.rb	mailto scheme support
 weakref.rb	weak reference class

lib/uri.rb

@@ -0,0 +1,41 @@
+#
+# $Id$
+#
+# Copyright (c) 2001 akira yamada <akira@ruby-lang.org>
+# You can redistribute it and/or modify it under the same term as Ruby.
+#
+
+=begin
+
+  Copyright (c) 2001 akira yamada <akira@ruby-lang.org>
+  You can redistribute it and/or modify it under the same term as Ruby.
+
+= URI - URI support for Ruby
+
+=end
+
+module URI
+  VERSION_CODE = '000904'.freeze
+  VERSION = VERSION_CODE.scan(/../).collect{|n| n.to_i}.join('.').freeze
+end
+
+=begin
+
+== Components
+
+  * ((<URI>)) Module
+  * ((<URI::Generic>)) Class
+  * ((<URI::FTP>)) Class
+  * ((<URI::HTTP>)) Class
+  * ((<URI::HTTPS>)) Class
+  * ((<URI::LDAP>)) Class
+  * ((<URI::MailTo>)) Class
+
+=end
+require 'uri/common'
+require 'uri/generic'
+require 'uri/ftp'
+require 'uri/http'
+require 'uri/https'
+require 'uri/ldap'
+require 'uri/mailto'

lib/uri/common.rb

@@ -0,0 +1,419 @@
+#
+# $Id$
+#
+# Copyright (c) 2001 akira yamada <akira@ruby-lang.org>
+# You can redistribute it and/or modify it under the same term as Ruby.
+#
+
+=begin
+
+== URI
+
+=end
+
+module URI
+  module REGEXP
+    module PATTERN
+      # RFC 2396 (URI Generic Syntax)
+      # RFC 2732 (IPv6 Literal Addresses in URL's)
+      # RFC 2373 (IPv6 Addressing Architecture)
+
+      # alpha         = lowalpha | upalpha
+      ALPHA = "a-zA-Z"
+      # alphanum      = alpha | digit
+      ALNUM = "#{ALPHA}\\d"
+
+      # hex           = digit | "A" | "B" | "C" | "D" | "E" | "F" |
+      #                         "a" | "b" | "c" | "d" | "e" | "f"
+      HEX     = "a-fA-F\\d"
+      # escaped       = "%" hex hex
+      ESCAPED = "%[#{HEX}]{2}"
+      # mark          = "-" | "_" | "." | "!" | "~" | "*" | "'" |
+      #                 "(" | ")"
+      # unreserved    = alphanum | mark
+      UNRESERVED = "-_.!~*'()#{ALNUM}"
+      # reserved      = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
+      #                 "$" | ","
+      # reserved      = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" | 
+      #                 "$" | "," | "[" | "]" (RFC 2732)
+      RESERVED = ";/?:@&=+$,\\[\\]"
+
+      # uric          = reserved | unreserved | escaped
+      URIC = "(?:[#{UNRESERVED}#{RESERVED}]|#{ESCAPED})"
+      # uric_no_slash = unreserved | escaped | ";" | "?" | ":" | "@" |
+      #                 "&" | "=" | "+" | "$" | ","
+      URIC_NO_SLASH = "(?:[#{UNRESERVED};?:@&=+$,]|#{ESCAPED})"
+      # query         = *uric
+      QUERY = "#{URIC}*"
+      # fragment      = *uric
+      FRAGMENT = "#{URIC}*"
+
+      # domainlabel   = alphanum | alphanum *( alphanum | "-" ) alphanum
+      DOMLABEL = "(?:[#{ALNUM}](?:[-#{ALNUM}]*[#{ALNUM}])?)"
+      # toplabel      = alpha | alpha *( alphanum | "-" ) alphanum
+      TOPLABEL = "(?:[#{ALPHA}](?:[-#{ALNUM}]*[#{ALNUM}])?)"
+      # hostname      = *( domainlabel "." ) toplabel [ "." ]
+      HOSTNAME = "(?:#{DOMLABEL}\\.)*#{TOPLABEL}\\.?"
+
+      # RFC 2373, APPENDIX B:
+      # IPv6address = hexpart [ ":" IPv4address ]
+      # IPv4address   = 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT
+      # hexpart = hexseq | hexseq "::" [ hexseq ] | "::" [ hexseq ]
+      # hexseq  = hex4 *( ":" hex4)
+      # hex4    = 1*4HEXDIG
+      #
+      # XXX: This definition has a flaw. "::" + IPv4address must be
+      # allowed too.  Here is a replacement.
+      #
+      # IPv4address = 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT
+      IPV4ADDR = "\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}"
+      # hex4     = 1*4HEXDIG
+      HEX4 = "[#{HEX}]{1,4}"
+      # lastpart = hex4 | IPv4address
+      LASTPART = "(?:#{HEX4}|#{IPV4ADDR})"
+      # hexseq1  = *( hex4 ":" ) hex4
+      HEXSEQ1 = "(?:#{HEX4}:)*#{HEX4}"
+      # hexseq2  = *( hex4 ":" ) lastpart
+      HEXSEQ2 = "(?:#{HEX4}:)*#{LASTPART}"
+      # IPv6address = hexseq2 | [ hexseq1 ] "::" [ hexseq2 ]
+      IPV6ADDR = "(?:#{HEXSEQ2}|(?:#{HEXSEQ1})?::(?:#{HEXSEQ2})?)"
+
+      # IPv6prefix  = ( hexseq1 | [ hexseq1 ] "::" [ hexseq1 ] ) "/" 1*2DIGIT
+      # unused
+
+      # ipv6reference = "[" IPv6address "]" (RFC 2732)
+      IPV6REF = "\\[#{IPV6ADDR}\\]"
+
+      # host          = hostname | IPv4address
+      # host          = hostname | IPv4address | IPv6reference (RFC 2732)
+      HOST = "(?:#{HOSTNAME}|#{IPV4ADDR}|#{IPV6REF})"
+      # port          = *digit
+      PORT = "\d*"
+      # hostport      = host [ ":" port ]
+      HOSTPORT = "#{HOST}(?:#{PORT})?"
+
+      # userinfo      = *( unreserved | escaped |
+      #                    ";" | ":" | "&" | "=" | "+" | "$" | "," )
+      USERINFO = "(?:[#{UNRESERVED};:&=+$,]|#{ESCAPED})*"
+
+      # pchar         = unreserved | escaped |
+      #                 ":" | "@" | "&" | "=" | "+" | "$" | ","
+      PCHAR = "(?:[#{UNRESERVED}:@&=+$,]|#{ESCAPED})"
+      # param         = *pchar
+      PARAM = "#{PCHAR}*"
+      # segment       = *pchar *( ";" param )
+      SEGMENT = "#{PCHAR}*(?:;#{PARAM})*"
+      # path_segments = segment *( "/" segment )
+      PATH_SEGMENTS = "#{SEGMENT}(?:/#{SEGMENT})*"
+
+      # server        = [ [ userinfo "@" ] hostport ]
+      SERVER = "(?:#{USERINFO}@)?#{HOSTPORT}"
+      # reg_name      = 1*( unreserved | escaped | "$" | "," |
+      #                     ";" | ":" | "@" | "&" | "=" | "+" )
+      REG_NAME = "(?:[#{UNRESERVED}$,;+@&=+]|#{ESCAPED})+"
+      # authority     = server | reg_name
+      AUTHORITY = "(?:#{SERVER}|#{REG_NAME})"
+
+      # rel_segment   = 1*( unreserved | escaped |
+      #                     ";" | "@" | "&" | "=" | "+" | "$" | "," )
+      REL_SEGMENT = "(?:[#{UNRESERVED};@&=+$,]|#{ESCAPED})+"
+
+      # scheme        = alpha *( alpha | digit | "+" | "-" | "." )
+      SCHEME = "[#{ALPHA}][-+.#{ALPHA}\\d]*"
+
+      # abs_path      = "/"  path_segments
+      ABS_PATH = "/#{PATH_SEGMENTS}"
+      # rel_path      = rel_segment [ abs_path ]
+      REL_PATH = "#{REL_SEGMENT}(?:#{ABS_PATH})?"
+      # net_path      = "//" authority [ abs_path ]
+      NET_PATH   = "//#{AUTHORITY}(?:#{ABS_PATH})?"
+
+      # hier_part     = ( net_path | abs_path ) [ "?" query ]
+      HIER_PART   = "(?:#{NET_PATH}|#{ABS_PATH})(?:\\?(?:#{QUERY}))?"
+      # opaque_part   = uric_no_slash *uric
+      OPAQUE_PART = "#{URIC_NO_SLASH}#{URIC}*"
+
+      # absoluteURI   = scheme ":" ( hier_part | opaque_part )
+      ABS_URI   = "#{SCHEME}:(?:#{HIER_PART}|#{OPAQUE_PART})"
+      # relativeURI   = ( net_path | abs_path | rel_path ) [ "?" query ]
+      REL_URI = "(?:#{NET_PATH}|#{ABS_PATH}|#{REL_PATH})(?:\\?#{QUERY})?"
+
+      # URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]
+      URI_REF = "(?:#{ABS_URI}|#{REL_URI})?(?:##{FRAGMENT})?"
+
+      # XXX:
+      X_ABS_URI = "
+	(#{PATTERN::SCHEME}):                     (?# 1: scheme)
+	(?:
+	   (?:
+	     //(?:
+		 (?:(?:(#{PATTERN::USERINFO})@)?  (?# 2: userinfo)
+		   (?:(#{PATTERN::HOST})(?::(\\d*))?))?(?# 3: host, 4: port)
+	       |
+		 (#{PATTERN::REG_NAME})           (?# 5: registry)
+	       )
+	       (#{PATTERN::ABS_PATH})?            (?# 6: path)
+	   )(?:\\?(#{PATTERN::QUERY}))?           (?# 7: query)
+	|
+	   (#{PATTERN::OPAQUE_PART})              (?# 8: opaque)
+	)
+	(?:\\#(#{PATTERN::FRAGMENT}))?            (?# 9: fragment)
+      "
+      X_REL_URI = "
+        (?:
+          (?:
+            //
+            (?:
+              (?:(#{PATTERN::USERINFO})@)?       (?# 1: userinfo)
+                (#{PATTERN::HOST})?(?::(\\d*))?  (?# 2: host, 3: port)
+            |
+              (#{PATTERN::REG_NAME})             (?# 4: registry)
+            )
+          )
+        |
+          (#{PATTERN::REL_SEGMENT})              (?# 5: rel_segment)
+        )?
+        (#{PATTERN::ABS_PATH})?                  (?# 6: abs_path)
+        (?:\\?(#{PATTERN::QUERY}))?              (?# 7: query)
+        (?:\\#(#{PATTERN::FRAGMENT}))?           (?# 8: fragment)
+      "
+    end # PATTERN
+
+    # for URI::split
+    ABS_URI = Regexp.new('^' + PATTERN::X_ABS_URI + '$', #'
+			 Regexp::EXTENDED, 'N').freeze
+    REL_URI = Regexp.new('^' + PATTERN::X_REL_URI + '$', #'
+			 Regexp::EXTENDED, 'N').freeze
+
+    # for URI::extract
+    URI_REF     = Regexp.new(PATTERN::URI_REF, false, 'N').freeze
+    ABS_URI_REF = Regexp.new(PATTERN::X_ABS_URI, Regexp::EXTENDED, 'N').freeze
+    REL_URI_REF = Regexp.new(PATTERN::X_REL_URI, Regexp::EXTENDED, 'N').freeze
+
+    # for URI::escape/unescape
+    ESCAPED = Regexp.new(PATTERN::ESCAPED, false, 'N').freeze
+    UNSAFE  = Regexp.new("[^#{PATTERN::UNRESERVED}#{PATTERN::RESERVED}]",
+			 false, 'N').freeze
+
+    # for Generic#initialize
+    SCHEME   = Regexp.new("^#{PATTERN::SCHEME}$", false, 'N').freeze #"
+    USERINFO = Regexp.new("^#{PATTERN::USERINFO}$", false, 'N').freeze #"
+    HOST     = Regexp.new("^#{PATTERN::HOST}$", false, 'N').freeze #"
+    PORT     = Regexp.new("^#{PATTERN::PORT}$", false, 'N').freeze #"
+    OPAQUE   = Regexp.new("^#{PATTERN::OPAQUE_PART}$", false, 'N').freeze #"
+    REGISTRY = Regexp.new("^#{PATTERN::REG_NAME}$", false, 'N').freeze #"
+    ABS_PATH = Regexp.new("^#{PATTERN::ABS_PATH}$", false, 'N').freeze #"
+    REL_PATH = Regexp.new("^#{PATTERN::REL_PATH}$", false, 'N').freeze #"
+    QUERY    = Regexp.new("^#{PATTERN::QUERY}$", false, 'N').freeze #"
+    FRAGMENT = Regexp.new("^#{PATTERN::FRAGMENT}$", false, 'N').freeze #"
+  end # REGEXP
+
+  module Util
+    def make_components_hash(klass, array_hash)
+      tmp = {}
+      if array_hash.kind_of?(Array) &&
+	  array_hash.size == klass.component.size - 1
+	klass.component[1..-1].each_index do |i|
+	  begin
+	    tmp[klass.component[i + 1]] = array_hash[i].clone
+	  rescue TypeError
+	    tmp[klass.component[i + 1]] = array_hash[i]
+	  end
+	end
+
+      elsif array_hash.kind_of?(Hash)
+	array_hash.each do |key, value|
+	  begin
+	    tmp[key] = value.clone
+	  rescue TypeError
+	    tmp[key] = value
+	  end
+	end
+      else
+	raise ArgumentError, 
+	  "expected Array of or Hash of compornents of #{klass.to_s} (#{klass.component[1..-1].join(', ')})"
+      end
+      tmp[:scheme] = klass.to_s.sub(/\A.*::/, '').downcase
+
+      return tmp
+    end
+    module_function :make_components_hash
+  end
+
+  module Escape
+    include REGEXP
+
+    def escape(str, unsafe = UNSAFE)
+      unless unsafe.kind_of?(Regexp)
+	# perhaps unsafe is String object
+	unsafe = Regexp.new(Regexp.quote(unsafe), false, 'N')
+      end
+      str.gsub(unsafe) do |us|
+	tmp = ''
+	us.each_byte do |uc|
+	  tmp << sprintf('%%%02X', uc)
+	end
+	tmp
+      end
+    end
+    alias encode escape
+
+    def unescape(str)
+      str.gsub(ESCAPED) do
+	$&[1,2].hex.chr
+      end
+    end
+    alias decode unescape
+  end
+
+  include REGEXP
+  extend Escape
+
+  @@schemes = {}
+
+  class Error < StandardError; end
+  class InvalidURIError < Error; end # it is not URI.
+  class InvalidComponentError < Error; end # it is not component of URI.
+  class BadURIError < Error; end # the URI is valid but it is bad for the position.
+
+=begin
+
+=== Methods
+
+--- URI::split(uri)
+
+=end
+
+  def self.split(uri)
+    case uri
+    when ''
+      # null uri
+
+    when ABS_URI
+      scheme, userinfo, host, port, 
+	registry, path, query, opaque, fragment = $~[1..-1]
+
+      # URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]
+
+      # absoluteURI   = scheme ":" ( hier_part | opaque_part )
+      # hier_part     = ( net_path | abs_path ) [ "?" query ]
+      # opaque_part   = uric_no_slash *uric
+
+      # abs_path      = "/"  path_segments
+      # net_path      = "//" authority [ abs_path ]
+
+      # authority     = server | reg_name
+      # server        = [ [ userinfo "@" ] hostport ]
+
+      if !scheme
+	raise InvalidURIError, 
+	  "bad URI(absolute but no scheme): #{uri}"
+      end
+      if !opaque && (!path && (!host && !registry))
+	raise InvalidURIError,
+	  "bad URI(absolute but no path): #{uri}" 
+      end
+
+    when REL_URI
+      scheme = nil
+      opaque = nil
+
+      userinfo, host, port, registry, 
+	rel_segment, abs_path, query, fragment = $~[1..-1]
+      if rel_segment && abs_path
+	path = rel_segment + abs_path
+      elsif rel_segment
+	path = rel_segment
+      elsif abs_path
+	path = abs_path
+      end
+
+      # URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]
+
+      # relativeURI   = ( net_path | abs_path | rel_path ) [ "?" query ]
+
+      # net_path      = "//" authority [ abs_path ]
+      # abs_path      = "/"  path_segments
+      # rel_path      = rel_segment [ abs_path ]
+
+      # authority     = server | reg_name
+      # server        = [ [ userinfo "@" ] hostport ]
+
+    else
+      raise InvalidURIError, "bad URI(is not URI?): #{uri}"
+    end
+
+    path = '' if !path && !opaque # (see RFC2396 Section 5.2)
+    ret = [
+      scheme, 
+      userinfo, host, port, 	# X
+      registry,			# X
+      path, 			# Y
+      opaque,			# Y
+      query,
+      fragment
+    ]
+    return ret
+  end
+
+=begin
+
+--- URI::parse(uri_str)
+
+=end
+  def self.parse(uri)
+    scheme, userinfo, host, port, 
+      registry, path, opaque, query, fragment = self.split(uri)
+
+    if scheme && @@schemes.include?(scheme.upcase)
+      @@schemes[scheme.upcase].new(scheme, userinfo, host, port, 
+				   registry, path, opaque, query, 
+				   fragment)
+    else
+      Generic.new(scheme, userinfo, host, port, 
+		  registry, path, opaque, query, 
+		  fragment)
+    end
+  end
+
+=begin
+
+--- URI::extract(str[, schemes])
+
+=end
+  def self.extract(str, schemes = [])
+    urls = []
+    if schemes.size > 0
+      tmp = Regexp.new('(?:' + schemes.collect{|s| 
+			 Regexp.quote(s + ':')
+		       }.join('|') + ')', 
+		       Regexp::IGNORECASE, 'N')
+      str.scan(tmp) {
+	tmp_str = $& + $'
+	if ABS_URI_REF =~ tmp_str
+	  if block_given?
+	    yield($&)
+	  else
+	    urls << $&
+	  end
+	end
+      }
+
+    else
+      str.scan(ABS_URI_REF) {
+	if block_given?
+	  yield($&)
+	else
+	  urls << $&
+	end
+      }
+    end
+
+    if block_given?
+      return nil
+    else
+      return urls
+    end
+  end
+
+end # URI

lib/uri/ftp.rb

@@ -0,0 +1,149 @@
+#
+# $Id$
+#
+# Copyright (c) 2001 akira yamada <akira@ruby-lang.org>
+# You can redistribute it and/or modify it under the same term as Ruby.
+#
+
+require 'uri/generic'
+
+module URI
+
+=begin
+
+== URI::FTP
+
+=== Super Class
+
+((<URI::Generic>))
+
+=end
+
+  # RFC1738 section 3.2.
+  class FTP < Generic
+    DEFAULT_PORT = 21
+
+    COMPONENT = [
+      :scheme, 
+      :userinfo, :host, :port,
+      :path, :typecode
+    ].freeze
+
+    TYPECODE = ['a', 'i', 'd'].freeze
+    TYPECODE_PREFIX = ';type='.freeze
+
+=begin
+
+=== Class Methods
+
+--- URI::FTP::build
+    Create a new URI::FTP object from components of URI::FTP with
+    check.  It is scheme, userinfo, host, port, path and typecode. It
+    provided by an Array or a Hash. typecode is "a", "i" or "d".
+
+--- URI::FTP::new
+    Create a new URI::FTP object from ``generic'' components with no
+    check.
+
+=end
+
+    def self.new2(user, password, host, port, path, 
+		  typecode = nil, arg_check = true)
+      typecode = nil if typecode.size == 0
+      if typecode && !TYPECODE.include?(typecode)
+	raise ArgumentError,
+	  "bad typecode is specified: #{typecode}"
+      end
+
+      # do escape
+
+      self.new('ftp',
+	       [user, password], 
+	       host, port, nil, 
+	       typecode ? path + TYPECODE_PREFIX + typecode : path, 
+	       nil, nil, nil, arg_check)
+    end
+
+    def self.build(args)
+      tmp = Util::make_components_hash(self, args)
+
+      if tmp[:typecode]
+	if tmp[:typecode].size == 1
+	  tmp[:typecode] = TYPECODE_PREFIX + tmp[:typecode] 
+	end
+	tmp[:path] << tmp[:typecode]
+      end
+
+      return super(tmp)
+    end
+
+    def initialize(*arg)
+      super(*arg)
+      @typecode = nil
+      tmp = @path.index(TYPECODE_PREFIX)
+      if tmp
+	typecode = @path[tmp + TYPECODE_PREFIX.size..-1]
+	self.set_path(@path[0..tmp - 1])
+	
+	if arg[-1]
+	  self.typecode = typecode
+	else
+	  self.set_typecode(typecode)
+	end
+      end
+    end
+    attr_reader :typecode
+
+    #
+    # methods for typecode
+    #
+
+    def check_typecode(v)
+      if TYPECODE.include?(v)
+	return true
+      else
+	raise InvalidComponentError,
+	  "bad typecode(expected #{TYPECODE.join(', ')}): #{v}"
+      end
+    end
+    private :check_typecode
+
+    def set_typecode(v)
+      @typecode = v
+    end
+    protected :set_typecode
+
+    def typecode=(typecode)
+      check_typecode(typecode)
+      set_typecode(typecode)
+    end
+
+=begin
+=end
+    def merge(oth)
+      tmp = super(oth)
+      if self != tmp
+	tmp.set_typecode(oth.typecode)
+      end
+
+      return tmp
+    end
+
+=begin
+=end
+    def to_str
+      save_path = nil
+      if @typecode
+	save_path = @path
+	@path = @path + TYPECODE_PREFIX + @typecode
+      end
+      str = super
+      if @typecode
+	@path = save_path
+      end
+
+      return str
+    end
+  end # FTP
+  @@schemes['FTP'] = FTP
+end # URI

lib/uri/http.rb

@@ -0,0 +1,76 @@
+#
+# $Id$
+#
+# Copyright (c) 2001 akira yamada <akira@ruby-lang.org>
+# You can redistribute it and/or modify it under the same term as Ruby.
+#
+
+require 'uri/generic'
+
+module URI
+
+=begin
+
+== URI::HTTP
+
+=== Super Class
+
+((<URI::Generic>))
+
+=end
+
+  # RFC1738 section 3.3.
+  class HTTP < Generic
+    DEFAULT_PORT = 80
+
+    COMPONENT = [
+      :scheme, 
+      :userinfo, :host, :port, 
+      :path, 
+      :query, 
+      :fragment
+    ].freeze
+
+=begin
+
+=== Class Methods
+
+--- URI::HTTP::build
+    Create a new URI::HTTP object from components of URI::HTTP with
+    check.  It is scheme, userinfo, host, port, path, query and
+    fragment. It provided by an Array of a Hash.
+
+--- URI::HTTP::new
+    Create a new URI::HTTP object from ``generic'' components with no
+    check.
+
+=end
+
+    def self.build(args)
+      tmp = Util::make_components_hash(self, args)
+      return super(tmp)
+    end
+
+    def initialize(*arg)
+      super(*arg)
+    end
+
+=begin
+
+=== Instance Methods
+
+--- URI::HTTP#request_uri
+
+=end
+    def request_uri
+      r = path_query
+      if r[0] != ?/
+	r = '/' + r
+      end
+
+      r
+    end
+  end # HTTP
+
+  @@schemes['HTTP'] = HTTP
+end # URI

lib/uri/https.rb

@@ -0,0 +1,26 @@
+#
+# $Id$
+#
+# Copyright (c) 2001 akira yamada <akira@ruby-lang.org>
+# You can redistribute it and/or modify it under the same term as Ruby.
+#
+
+require 'uri/http'
+
+module URI
+
+=begin
+
+== URI::HTTPS
+
+=== Super Class
+
+((<URI::HTTP>))
+
+=end
+
+  class HTTPS < HTTP
+    DEFAULT_PORT = 443
+  end
+  @@schemes['HTTPS'] = HTTPS
+end # URI

lib/uri/ldap.rb

@@ -0,0 +1,238 @@
+#
+# $Id$
+#
+
+require 'uri/generic'
+
+module URI
+
+=begin
+
+== URI::LDAP
+
+URI::LDAP is copyrighted free software by Takaaki Tateishi and akira yamada.
+
+  Copyright (c) 2001 Takaaki Tateishi <ttate@jaist.ac.jp> and 
+  akira yamada <akira@ruby-lang.org>.
+  You can redistribute it and/or modify it under the same term as Ruby.
+
+=== Super Class
+
+((<URI::Generic>))
+
+=end
+
+  # LDAP URI SCHEMA (described in RFC2255)
+  # ldap://<host>/<dn>[?<attrs>[?<scope>[?<filter>[?<extensions>]]]]
+  class LDAP < Generic
+
+    DEFAULT_PORT = 389
+    
+    COMPONENT = [
+      :scheme,
+      :host, :port,
+      :dn,
+      :attributes,
+      :scope,
+      :filter,
+      :extensions,
+    ].freeze
+
+    SCOPE = [
+      SCOPE_ONE = 'one',
+      SCOPE_SUB = 'sub',
+      SCOPE_BASE = 'base',
+    ].freeze
+
+=begin
+
+=== Class Methods
+
+--- URI::LDAP::build
+
+--- URI::LDAP::new
+
+=end
+
+    def self.build(args)
+      tmp = Util::make_components_hash(self, args)
+
+      if tmp[:dn]
+	tmp[:path] = tmp[:dn]
+      end
+
+      query = []
+      [:extensions, :filter, :scope, :attributes].collect do |x|
+	next if !tmp[x] && query.size == 0
+	query.unshift(tmp[x])
+      end
+
+      tmp[:query] = query.join('?')
+
+      return super(tmp)
+    end
+
+    def initialize(*arg)
+      super(*arg)
+
+      if @fragment
+	raise InvalidURIError, 'bad LDAP URL'
+      end
+
+      parse_dn
+      parse_query
+    end
+
+    def parse_dn
+      @dn = @path[1..-1]
+    end
+    private :parse_dn
+
+    def parse_query
+      @attributes = nil
+      @scope      = nil
+      @filter     = nil
+      @extensions = nil
+
+      if @query
+	attrs, scope, filter, extensions = @query.split('?')
+
+	@attributes = attrs if attrs && attrs.size > 0
+	@scope      = scope if scope && scope.size > 0
+	@filter     = filter if filter && filter.size > 0
+	@extensions = extensions if extensions && extensions.size > 0
+      end
+    end
+    private :parse_query
+
+    def build_path_query
+      @path = '/' + @dn
+
+      query = []
+      [@extensions, @filter, @scope, @attributes].each do |x|
+	next if !x && query.size == 0
+	query.unshift(x)
+      end
+      @query = query.join('?')
+    end
+    private :build_path_query
+
+=begin
+
+=== Instance Methods
+
+--- URI::LDAP#dn
+
+--- URI::LDAP#dn=(v)
+
+=end
+
+    def dn
+      @dn
+    end
+
+    def set_dn(val)
+      @dn = val
+      build_path_query
+    end
+    protected :set_dn
+
+    def dn=(val)
+      set_dn(val)
+    end
+
+=begin
+
+--- URI::LDAP#attributes
+
+--- URI::LDAP#attributes=(v)
+
+=end
+
+    def attributes
+      @attributes
+    end
+
+    def set_attributes(val)
+      @attributes = val
+      build_path_query
+    end
+    protected :set_attributes
+
+    def attributes=(val)
+      set_attributes(val)
+    end
+
+=begin
+
+--- URI::LDAP#scope
+
+--- URI::LDAP#scope=(v)
+
+=end
+
+    def scope
+      @scope
+    end
+
+    def set_scope(val)
+      @scope = val
+      build_path_query
+    end
+    protected :set_scope
+
+    def scope=(val)
+      set_scope(val)
+    end
+
+=begin
+
+--- URI::LDAP#filter
+
+--- URI::LDAP#filter=(v)
+
+=end
+
+    def filter
+      @filter
+    end
+
+    def set_filter(val)
+      @filter = val
+      build_path_query
+    end
+    protected :set_filter
+
+    def filter=(val)
+      set_filter(val)
+    end
+
+=begin
+
+--- URI::LDAP#extensions
+
+--- URI::LDAP#extensions=(v)
+
+=end
+
+    def extensions
+      @extensions
+    end
+
+    def set_extensions(val)
+      @extensions = val
+      build_path_query
+    end
+    protected :set_extensions
+
+    def extensions=(val)
+      set_extensions(val)
+    end
+  end
+
+  def hierarchical?
+    false
+  end
+
+  @@schemes['LDAP'] = LDAP
+end

lib/uri/mailto.rb

@@ -0,0 +1,260 @@
+#
+# $Id$
+#
+# Copyright (c) 2001 akira yamada <akira@ruby-lang.org>
+# You can redistribute it and/or modify it under the same term as Ruby.
+#
+
+require 'uri/generic'
+
+module URI
+
+=begin
+
+== URI::MailTo
+
+=== Super Class
+
+((<URI::Generic>))
+
+=end
+
+  # RFC2368, The mailto URL scheme
+  class MailTo < Generic
+    include REGEXP
+
+    DEFAULT_PORT = nil
+
+    COMPONENT = [
+      :scheme,
+      :to, :headers
+    ].freeze
+
+    #  "hname" and "hvalue" are encodings of an RFC 822 header name and
+    #  value, respectively. As with "to", all URL reserved characters must
+    #  be encoded.
+    #
+    #  "#mailbox" is as specified in RFC 822 [RFC822]. This means that it
+    #  consists of zero or more comma-separated mail addresses, possibly
+    #  including "phrase" and "comment" components. Note that all URL
+    #  reserved characters in "to" must be encoded: in particular,
+    #  parentheses, commas, and the percent sign ("%"), which commonly occur
+    #  in the "mailbox" syntax.
+    #
+    #  Within mailto URLs, the characters "?", "=", "&" are reserved.
+
+    # hname      =  *urlc
+    # hvalue     =  *urlc
+    # header     =  hname "=" hvalue
+    HEADER_REGEXP = "(?:[^?=&]*=[^?=&]*)".freeze
+    # headers    =  "?" header *( "&" header )
+    # to         =  #mailbox
+    # mailtoURL  =  "mailto:" [ to ] [ headers ]
+    MAILBOX_REGEXP = "(?:[^(),%?=&]|#{PATTERN::ESCAPED})".freeze
+    MAILTO_REGEXP = Regexp.new("
+      \\A
+      (#{MAILBOX_REGEXP}*?)                         (?# 1: to)
+      (?:
+        \\?
+        (#{HEADER_REGEXP}(?:\\&#{HEADER_REGEXP})*)  (?# 2: headers)
+      )?
+      \\z
+    ", Regexp::EXTENDED, 'N').freeze
+
+=begin
+
+=== Class Methods
+
+--- URI::MailTo::build
+    Create a new URI::MailTo object from components of URI::MailTo
+    with check.  It is to and headers. It provided by an Array of a
+    Hash. You can provide headers as an String like
+    "subject=subscribe&cc=addr" or an Array like [["subject",
+    "subscribe"], ["cc", "addr"]]
+
+--- URI::MailTo::new
+    Create a new URI::MailTo object from ``generic'' components with
+    no check. Because, this method is usually called from URI::parse
+    and the method checks validity of each components.
+
+=end
+
+    def self.build(args)
+      tmp = Util::make_components_hash(self, args)
+
+      if tmp[:to]
+	tmp[:opaque] = tmp[:to]
+      else
+	tmp[:opaque] = ''
+      end
+
+      if tmp[:headers]
+	tmp[:opaque] << '?'
+
+	if tmp[:headers].kind_of?(Array)
+	  tmp[:opaque] << tmp[:headers].collect { |x|
+	    if x.kind_of?(Array)
+	      x[0] + '=' + x[1..-1].to_s
+	    else
+	      x.to_s
+	    end
+	  }.join('&')
+
+	elsif tmp[:headers].kind_of?(Hash)
+	  tmp[:opaque] << tmp[:headers].collect { |h,v|
+	    h + '=' + v
+	  }.join('&')
+
+	else
+	  tmp[:opaque] << tmp[:headers].to_s
+	end
+      end
+
+      return super(tmp)
+    end
+
+    def initialize(*arg)
+      super(*arg)
+
+      @to = nil
+      @headers = []
+
+      if MAILTO_REGEXP =~ @opaque
+ 	if arg[-1]
+	  self.to = $1
+	  self.headers = $2
+	else
+	  set_to($1)
+	  set_headers($2)
+	end
+      elsif arg[-1]
+	raise InvalidComponentError,
+	  "unrecognised opaque part for mailtoURL: #{@opaque}"
+      end
+    end
+    attr_reader :to
+    attr_reader :headers
+
+=begin
+
+=== Instance Methods
+
+--- URI::MailTo#to
+
+--- URI::MailTo#to=(v)
+
+=end
+
+    #
+    # methods for to
+    #
+
+    def check_to(v)
+      return true unless v
+      return true if v.size == 0
+
+      if OPAQUE !~ v || /\A#{MAILBOX_REGEXP}*\z/o !~ v
+	raise InvalidComponentError,
+	  "bad component(expected opaque component): #{v}"
+      end
+
+      return true
+    end
+    private :check_to
+
+    def set_to(v)
+      @to = v
+    end
+    protected :set_to
+
+    def to=(v)
+      check_to(v)
+      set_to(v)
+    end
+
+=begin
+
+--- URI::MailTo#headers
+
+--- URI::MailTo#headers=(v)
+
+=end
+
+    #
+    # methods for headers
+    #
+
+    def check_headers(v)
+      return true unless v
+      return true if v.size == 0
+
+      if OPAQUE !~ v || 
+	  /\A(#{HEADER_REGEXP}(?:\&#{HEADER_REGEXP})*)\z/o !~ v
+	raise InvalidComponentError,
+	  "bad component(expected opaque component): #{v}"
+      end
+
+      return true
+    end
+    private :check_headers
+
+    def set_headers(v)
+      @headers = []
+      if v
+	v.scan(HEADER_REGEXP) do |x|
+	  @headers << x.split(/=/o, 2)
+	end
+      end
+    end
+    protected :set_headers
+
+    def headers=(v)
+      check_headers(v)
+      set_headers(v)
+    end
+
+    def to_str
+      @scheme + ':' + 
+	if @to 
+	  @to
+	else
+	  ''
+	end + 
+	if @headers.size > 0
+	  '?' + @headers.collect{|x| x.join('=')}.join('&')
+	else
+	  ''
+	end
+    end
+
+=begin
+
+--- URI::MailTo#to_mailtext
+
+=end
+    def to_mailtext
+      to = URI::unescape(@to)
+      head = ''
+      body = ''
+      @headers.each do |x|
+	case x[0]
+	when 'body'
+	  body = URI::unescape(x[1])
+	when 'to'
+	  to << ', ' + URI::unescape(x[1])
+	else
+	  head << URI::unescape(x[0]).capitalize + ': ' +
+	    URI::unescape(x[1])  + "\n"
+	end
+      end
+
+      return "To: #{to}
+#{head}
+#{body}
+"
+    end
+    alias to_rfc822text to_mailtext
+  end # MailTo
+
+  @@schemes['MAILTO'] = MailTo
+end # URI