Server IP : 66.29.132.122 / Your IP : 3.142.12.31 Web Server : LiteSpeed System : Linux business142.web-hosting.com 4.18.0-553.lve.el8.x86_64 #1 SMP Mon May 27 15:27:34 UTC 2024 x86_64 User : admazpex ( 531) PHP Version : 7.2.34 Disable Function : NONE MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : ON | Sudo : OFF | Pkexec : OFF Directory : /proc/self/root/proc/thread-self/root/proc/thread-self/root/opt/alt/ruby23/lib64/ruby/2.3.0/uri/ |
Upload File : |
# frozen_string_literal: false #-- # = uri/common.rb # # Author:: Akira Yamada <akira@ruby-lang.org> # Revision:: $Id$ # License:: # You can redistribute it and/or modify it under the same term as Ruby. # # See URI for general documentation # module URI # # Includes URI::REGEXP::PATTERN # module RFC2396_REGEXP # # Patterns used to parse URI's # module PATTERN # :stopdoc: # 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 = ";/?:@&=+$,\\[\\]" # domainlabel = alphanum | alphanum *( alphanum | "-" ) alphanum DOMLABEL = "(?:[#{ALNUM}](?:[-#{ALNUM}]*[#{ALNUM}])?)" # toplabel = alpha | alpha *( alphanum | "-" ) alphanum TOPLABEL = "(?:[#{ALPHA}](?:[-#{ALNUM}]*[#{ALNUM}])?)" # hostname = *( domainlabel "." ) toplabel [ "." ] HOSTNAME = "(?:#{DOMLABEL}\\.)*#{TOPLABEL}\\.?" # :startdoc: end # PATTERN # :startdoc: end # REGEXP # class that Parses String's into URI's # # It contains a Hash set of patterns and Regexp's that match and validate. # class RFC2396_Parser include RFC2396_REGEXP # # == Synopsis # # URI::Parser.new([opts]) # # == Args # # The constructor accepts a hash as options for parser. # Keys of options are pattern names of URI components # and values of options are pattern strings. # The constructor generates set of regexps for parsing URIs. # # You can use the following keys: # # * :ESCAPED (URI::PATTERN::ESCAPED in default) # * :UNRESERVED (URI::PATTERN::UNRESERVED in default) # * :DOMLABEL (URI::PATTERN::DOMLABEL in default) # * :TOPLABEL (URI::PATTERN::TOPLABEL in default) # * :HOSTNAME (URI::PATTERN::HOSTNAME in default) # # == Examples # # p = URI::Parser.new(:ESCAPED => "(?:%[a-fA-F0-9]{2}|%u[a-fA-F0-9]{4})") # u = p.parse("http://example.jp/%uABCD") #=> #<URI::HTTP:0xb78cf4f8 URL:http://example.jp/%uABCD> # URI.parse(u.to_s) #=> raises URI::InvalidURIError # # s = "http://example.com/ABCD" # u1 = p.parse(s) #=> #<URI::HTTP:0xb78c3220 URL:http://example.com/ABCD> # u2 = URI.parse(s) #=> #<URI::HTTP:0xb78b6d54 URL:http://example.com/ABCD> # u1 == u2 #=> true # u1.eql?(u2) #=> false # def initialize(opts = {}) @pattern = initialize_pattern(opts) @pattern.each_value(&:freeze) @pattern.freeze @regexp = initialize_regexp(@pattern) @regexp.each_value(&:freeze) @regexp.freeze end # The Hash of patterns. # # see also URI::Parser.initialize_pattern attr_reader :pattern # The Hash of Regexp # # see also URI::Parser.initialize_regexp attr_reader :regexp # Returns a split URI against regexp[:ABS_URI] def split(uri) case uri when '' # null uri when @regexp[:ABS_URI] scheme, opaque, userinfo, host, port, registry, path, query, 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 @regexp[: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 # # == Args # # +uri+:: # String # # == Description # # parses +uri+ and constructs either matching URI scheme object # (FTP, HTTP, HTTPS, LDAP, LDAPS, or MailTo) or URI::Generic # # == Usage # # p = URI::Parser.new # p.parse("ldap://ldap.example.com/dc=example?user=john") # #=> #<URI::LDAP:0x00000000b9e7e8 URL:ldap://ldap.example.com/dc=example?user=john> # def parse(uri) scheme, userinfo, host, port, registry, path, opaque, query, fragment = self.split(uri) if scheme && URI.scheme_list.include?(scheme.upcase) URI.scheme_list[scheme.upcase].new(scheme, userinfo, host, port, registry, path, opaque, query, fragment, self) else Generic.new(scheme, userinfo, host, port, registry, path, opaque, query, fragment, self) end end # # == Args # # +uris+:: # an Array of Strings # # == Description # # Attempts to parse and merge a set of URIs # def join(*uris) uris[0] = convert_to_uri(uris[0]) uris.inject :merge end # # :call-seq: # extract( str ) # extract( str, schemes ) # extract( str, schemes ) {|item| block } # # == Args # # +str+:: # String to search # +schemes+:: # Patterns to apply to +str+ # # == Description # # Attempts to parse and merge a set of URIs # If no +block+ given , then returns the result, # else it calls +block+ for each element in result. # # see also URI::Parser.make_regexp # def extract(str, schemes = nil) if block_given? str.scan(make_regexp(schemes)) { yield $& } nil else result = [] str.scan(make_regexp(schemes)) { result.push $& } result end end # returns Regexp that is default self.regexp[:ABS_URI_REF], # unless +schemes+ is provided. Then it is a Regexp.union with self.pattern[:X_ABS_URI] def make_regexp(schemes = nil) unless schemes @regexp[:ABS_URI_REF] else /(?=#{Regexp.union(*schemes)}:)#{@pattern[:X_ABS_URI]}/x end end # # :call-seq: # escape( str ) # escape( str, unsafe ) # # == Args # # +str+:: # String to make safe # +unsafe+:: # Regexp to apply. Defaults to self.regexp[:UNSAFE] # # == Description # # constructs a safe String from +str+, removing unsafe characters, # replacing them with codes. # def escape(str, unsafe = @regexp[:UNSAFE]) unless unsafe.kind_of?(Regexp) # perhaps unsafe is String object unsafe = Regexp.new("[#{Regexp.quote(unsafe)}]", false) end str.gsub(unsafe) do us = $& tmp = '' us.each_byte do |uc| tmp << sprintf('%%%02X', uc) end tmp end.force_encoding(Encoding::US_ASCII) end # # :call-seq: # unescape( str ) # unescape( str, unsafe ) # # == Args # # +str+:: # String to remove escapes from # +unsafe+:: # Regexp to apply. Defaults to self.regexp[:ESCAPED] # # == Description # # Removes escapes from +str+ # def unescape(str, escaped = @regexp[:ESCAPED]) str.gsub(escaped) { [$&[1, 2].hex].pack('C') }.force_encoding(str.encoding) end @@to_s = Kernel.instance_method(:to_s) def inspect @@to_s.bind(self).call end private # Constructs the default Hash of patterns def initialize_pattern(opts = {}) ret = {} ret[:ESCAPED] = escaped = (opts.delete(:ESCAPED) || PATTERN::ESCAPED) ret[:UNRESERVED] = unreserved = opts.delete(:UNRESERVED) || PATTERN::UNRESERVED ret[:RESERVED] = reserved = opts.delete(:RESERVED) || PATTERN::RESERVED ret[:DOMLABEL] = opts.delete(:DOMLABEL) || PATTERN::DOMLABEL ret[:TOPLABEL] = opts.delete(:TOPLABEL) || PATTERN::TOPLABEL ret[:HOSTNAME] = hostname = opts.delete(:HOSTNAME) # RFC 2396 (URI Generic Syntax) # RFC 2732 (IPv6 Literal Addresses in URL's) # RFC 2373 (IPv6 Addressing Architecture) # uric = reserved | unreserved | escaped ret[:URIC] = uric = "(?:[#{unreserved}#{reserved}]|#{escaped})" # uric_no_slash = unreserved | escaped | ";" | "?" | ":" | "@" | # "&" | "=" | "+" | "$" | "," ret[:URIC_NO_SLASH] = uric_no_slash = "(?:[#{unreserved};?:@&=+$,]|#{escaped})" # query = *uric ret[:QUERY] = query = "#{uric}*" # fragment = *uric ret[:FRAGMENT] = fragment = "#{uric}*" # hostname = *( domainlabel "." ) toplabel [ "." ] # reg-name = *( unreserved / pct-encoded / sub-delims ) # RFC3986 unless hostname ret[:HOSTNAME] = hostname = "(?:[a-zA-Z0-9\\-.]|%\\h\\h)+" end # 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 ret[:IPV4ADDR] = ipv4addr = "\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}" # hex4 = 1*4HEXDIG hex4 = "[#{PATTERN::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 ] ret[:IPV6ADDR] = ipv6addr = "(?:#{hexseq2}|(?:#{hexseq1})?::(?:#{hexseq2})?)" # IPv6prefix = ( hexseq1 | [ hexseq1 ] "::" [ hexseq1 ] ) "/" 1*2DIGIT # unused # ipv6reference = "[" IPv6address "]" (RFC 2732) ret[:IPV6REF] = ipv6ref = "\\[#{ipv6addr}\\]" # host = hostname | IPv4address # host = hostname | IPv4address | IPv6reference (RFC 2732) ret[:HOST] = host = "(?:#{hostname}|#{ipv4addr}|#{ipv6ref})" # port = *digit ret[:PORT] = port = '\d*' # hostport = host [ ":" port ] ret[:HOSTPORT] = hostport = "#{host}(?::#{port})?" # userinfo = *( unreserved | escaped | # ";" | ":" | "&" | "=" | "+" | "$" | "," ) ret[:USERINFO] = userinfo = "(?:[#{unreserved};:&=+$,]|#{escaped})*" # pchar = unreserved | escaped | # ":" | "@" | "&" | "=" | "+" | "$" | "," pchar = "(?:[#{unreserved}:@&=+$,]|#{escaped})" # param = *pchar param = "#{pchar}*" # segment = *pchar *( ";" param ) segment = "#{pchar}*(?:;#{param})*" # path_segments = segment *( "/" segment ) ret[:PATH_SEGMENTS] = path_segments = "#{segment}(?:/#{segment})*" # server = [ [ userinfo "@" ] hostport ] server = "(?:#{userinfo}@)?#{hostport}" # reg_name = 1*( unreserved | escaped | "$" | "," | # ";" | ":" | "@" | "&" | "=" | "+" ) ret[:REG_NAME] = reg_name = "(?:[#{unreserved}$,;:@&=+]|#{escaped})+" # authority = server | reg_name authority = "(?:#{server}|#{reg_name})" # rel_segment = 1*( unreserved | escaped | # ";" | "@" | "&" | "=" | "+" | "$" | "," ) ret[:REL_SEGMENT] = rel_segment = "(?:[#{unreserved};@&=+$,]|#{escaped})+" # scheme = alpha *( alpha | digit | "+" | "-" | "." ) ret[:SCHEME] = scheme = "[#{PATTERN::ALPHA}][\\-+.#{PATTERN::ALPHA}\\d]*" # abs_path = "/" path_segments ret[:ABS_PATH] = abs_path = "/#{path_segments}" # rel_path = rel_segment [ abs_path ] ret[:REL_PATH] = rel_path = "#{rel_segment}(?:#{abs_path})?" # net_path = "//" authority [ abs_path ] ret[:NET_PATH] = net_path = "//#{authority}(?:#{abs_path})?" # hier_part = ( net_path | abs_path ) [ "?" query ] ret[:HIER_PART] = hier_part = "(?:#{net_path}|#{abs_path})(?:\\?(?:#{query}))?" # opaque_part = uric_no_slash *uric ret[:OPAQUE_PART] = opaque_part = "#{uric_no_slash}#{uric}*" # absoluteURI = scheme ":" ( hier_part | opaque_part ) ret[:ABS_URI] = abs_uri = "#{scheme}:(?:#{hier_part}|#{opaque_part})" # relativeURI = ( net_path | abs_path | rel_path ) [ "?" query ] ret[:REL_URI] = rel_uri = "(?:#{net_path}|#{abs_path}|#{rel_path})(?:\\?#{query})?" # URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ] ret[:URI_REF] = "(?:#{abs_uri}|#{rel_uri})?(?:##{fragment})?" ret[:X_ABS_URI] = " (#{scheme}): (?# 1: scheme) (?: (#{opaque_part}) (?# 2: opaque) | (?:(?: //(?: (?:(?:(#{userinfo})@)? (?# 3: userinfo) (?:(#{host})(?::(\\d*))?))? (?# 4: host, 5: port) | (#{reg_name}) (?# 6: registry) ) | (?!//)) (?# XXX: '//' is the mark for hostport) (#{abs_path})? (?# 7: path) )(?:\\?(#{query}))? (?# 8: query) ) (?:\\#(#{fragment}))? (?# 9: fragment) " ret[:X_REL_URI] = " (?: (?: // (?: (?:(#{userinfo})@)? (?# 1: userinfo) (#{host})?(?::(\\d*))? (?# 2: host, 3: port) | (#{reg_name}) (?# 4: registry) ) ) | (#{rel_segment}) (?# 5: rel_segment) )? (#{abs_path})? (?# 6: abs_path) (?:\\?(#{query}))? (?# 7: query) (?:\\#(#{fragment}))? (?# 8: fragment) " ret end # Constructs the default Hash of Regexp's def initialize_regexp(pattern) ret = {} # for URI::split ret[:ABS_URI] = Regexp.new('\A\s*' + pattern[:X_ABS_URI] + '\s*\z', Regexp::EXTENDED) ret[:REL_URI] = Regexp.new('\A\s*' + pattern[:X_REL_URI] + '\s*\z', Regexp::EXTENDED) # for URI::extract ret[:URI_REF] = Regexp.new(pattern[:URI_REF]) ret[:ABS_URI_REF] = Regexp.new(pattern[:X_ABS_URI], Regexp::EXTENDED) ret[:REL_URI_REF] = Regexp.new(pattern[:X_REL_URI], Regexp::EXTENDED) # for URI::escape/unescape ret[:ESCAPED] = Regexp.new(pattern[:ESCAPED]) ret[:UNSAFE] = Regexp.new("[^#{pattern[:UNRESERVED]}#{pattern[:RESERVED]}]") # for Generic#initialize ret[:SCHEME] = Regexp.new("\\A#{pattern[:SCHEME]}\\z") ret[:USERINFO] = Regexp.new("\\A#{pattern[:USERINFO]}\\z") ret[:HOST] = Regexp.new("\\A#{pattern[:HOST]}\\z") ret[:PORT] = Regexp.new("\\A#{pattern[:PORT]}\\z") ret[:OPAQUE] = Regexp.new("\\A#{pattern[:OPAQUE_PART]}\\z") ret[:REGISTRY] = Regexp.new("\\A#{pattern[:REG_NAME]}\\z") ret[:ABS_PATH] = Regexp.new("\\A#{pattern[:ABS_PATH]}\\z") ret[:REL_PATH] = Regexp.new("\\A#{pattern[:REL_PATH]}\\z") ret[:QUERY] = Regexp.new("\\A#{pattern[:QUERY]}\\z") ret[:FRAGMENT] = Regexp.new("\\A#{pattern[:FRAGMENT]}\\z") ret end def convert_to_uri(uri) if uri.is_a?(URI::Generic) uri elsif uri = String.try_convert(uri) parse(uri) else raise ArgumentError, "bad argument (expected URI object or URI string)" end end end # class Parser end # module URI