cookiejar 0.2.9

data/lib/cookiejar.rb

@@ -0,0 +1,2 @@
+require 'cookiejar/cookie'
+require 'cookiejar/jar'

data/lib/cookiejar/cookie.rb

@@ -0,0 +1,252 @@
+require 'time'
+require 'uri'
+require 'cookiejar/cookie_validation'
+
+module CookieJar
+  
+  # Cookie is an immutable object which defines the data model of a HTTP Cookie.
+  # The data values within the cookie may be different from the
+  # values described in the literal cookie declaration.
+  # Specifically, the 'domain' and 'path' values may be set to defaults 
+  # based on the requested resource that resulted in the cookie being set.
+  class Cookie
+  
+    # [String] The name of the cookie.
+    attr_reader :name
+    # [String] The value of the cookie, without any attempts at decoding.
+    attr_reader :value
+    
+    # [String] The domain scope of the cookie. Follows the RFC 2965 
+    # 'effective host' rules. A 'dot' prefix indicates that it applies both
+    # to the non-dotted domain and child domains, while no prefix indicates
+    # that only exact matches of the domain are in scope.
+    attr_reader :domain
+    
+    # [String] The path scope of the cookie. The cookie applies to URI paths
+    # that prefix match this value.
+    attr_reader :path
+
+    # [Boolean] The secure flag is set to indicate that the cookie should
+    # only be sent securely. Nearly all HTTP User Agent implementations assume
+    # this to mean that the cookie should only be sent over a
+    # SSL/TLS-protected connection
+    attr_reader :secure
+
+    # [Boolean] Popular browser extension to mark a cookie as invisible
+    # to code running within the browser, such as JavaScript
+    attr_reader :http_only
+    
+    # [Fixnum] Version indicator, currently either
+    # * 0 for netscape cookies
+    # * 1 for RFC 2965 cookies
+    attr_reader :version
+    # [String] RFC 2965 field for indicating comment (or a location)
+    # describing the cookie to a usesr agent.
+    attr_reader :comment, :comment_url
+    # [Boolean] RFC 2965 field for indicating session lifetime for a cookie 
+    attr_reader :discard
+    # [Array<FixNum>, nil] RFC 2965 port scope for the cookie. If not nil,
+    # indicates specific ports on the HTTP server which should receive this
+    # cookie if contacted. 
+    attr_reader :ports
+    # [Time] Time when this cookie was first evaluated and created.
+    attr_reader :created_at
+
+    # Evaluate when this cookie will expire. Uses the original cookie fields
+    # for a max age or expires
+    #
+    # @return [Time, nil] Time of expiry, if this cookie has an expiry set
+    def expires_at
+      if @expiry.nil? || @expiry.is_a?(Time)
+        @expiry
+      else
+        @created_at + @expiry
+      end
+    end
+    
+    # Indicates whether the cookie is currently considered valid
+    #
+    # @param [Time] time to compare against, or 'now' if omitted
+    # @return [Boolean]
+    def expired? (time = Time.now)
+      expires_at != nil && time > expires_at
+    end
+
+    # Indicates whether the cookie will be considered invalid after the end
+    # of the current user session
+    # @return [Boolean] 
+    def session?
+      @expiry == nil || @discard
+    end
+    
+    # Create a cookie based on an absolute URI and the string value of a
+    # 'Set-Cookie' header.
+    #
+    # @param request_uri [String, URI] HTTP/HTTPS absolute URI of request.
+    # This is used to fill in domain and port if missing from the cookie,
+    # and to perform appropriate validation.
+    # @param set_cookie_value [String] HTTP value for the Set-Cookie header.
+    # @return [Cookie] created from the header string and request URI
+    # @raise [InvalidCookieError] on validation failure(s)
+    def self.from_set_cookie request_uri, set_cookie_value 
+      args = CookieJar::CookieValidation.parse_set_cookie set_cookie_value
+      args[:domain] = CookieJar::CookieValidation.determine_cookie_domain request_uri, args[:domain]
+      args[:path] = CookieJar::CookieValidation.determine_cookie_path request_uri, args[:path]
+      cookie = Cookie.new args
+      CookieJar::CookieValidation.validate_cookie request_uri, cookie
+      cookie
+    end
+
+    # Create a cookie based on an absolute URI and the string value of a
+    # 'Set-Cookie2' header.
+    #
+    # @param request_uri [String, URI] HTTP/HTTPS absolute URI of request.
+    # This is used to fill in domain and port if missing from the cookie,
+    # and to perform appropriate validation.
+    # @param set_cookie_value [String] HTTP value for the Set-Cookie2 header.
+    # @return [Cookie] created from the header string and request URI
+    # @raise [InvalidCookieError] on validation failure(s)
+    def self.from_set_cookie2 request_uri, set_cookie_value 
+      args = CookieJar::CookieValidation.parse_set_cookie2 set_cookie_value
+      args[:domain] = CookieJar::CookieValidation.determine_cookie_domain request_uri, args[:domain]
+      args[:path] = CookieJar::CookieValidation.determine_cookie_path request_uri, args[:path]
+      cookie = Cookie.new args
+      CookieJar::CookieValidation.validate_cookie request_uri, cookie
+      cookie
+    end
+    
+    # Returns cookie in a format appropriate to send to a server.
+    #
+    # @param [FixNum] 0 version, 0 for Netscape-style cookies, 1 for 
+    #   RFC2965-style.
+    # @param [Boolean] true prefix, for RFC2965, whether to prefix with
+    # "$Version=<version>;". Ignored for Netscape-style cookies
+    def to_s ver=0, prefix=true
+      case ver
+      when 0
+        "#{name}=#{value}"
+      when 1
+        # we do not need to encode path; the only characters required to be
+        # quoted must be escaped in URI
+        str = prefix ? "$Version=#{version};" : ""
+        str << "#{name}=#{value};$Path=\"#{path}\""
+        if domain.start_with? '.'
+          str << ";$Domain=#{domain}"
+        end
+        if ports
+          str << ";$Port=\"#{ports.join ','}\""
+        end
+        str
+      end
+    end
+
+    # Determine if a cookie should be sent given a request URI along with
+    # other options.
+    #
+    # This currently ignores domain.
+    #
+    # @param uri [String, URI] the requested page which may need to receive
+    # this cookie
+    # @param script [Boolean] indicates that cookies with the 'httponly'
+    # extension should be ignored
+    # @return [Boolean] whether this cookie should be sent to the server
+    def should_send? request_uri, script
+      uri = CookieJar::CookieValidation.to_uri request_uri
+      # cookie path must start with the uri, it must not be a secure cookie
+      # being sent over http, and it must not be a http_only cookie sent to
+      # a script
+      path_match   = uri.path.start_with? @path
+      secure_match = !(@secure && uri.scheme == 'http') 
+      script_match = !(script && @http_only)
+      expiry_match = !expired?
+      ports_match = ports.nil? || (ports.include? uri.port)
+      path_match && secure_match && script_match && expiry_match && ports_match
+    end
+    
+    def decoded_value
+      CookieJar::CookieValidation::decode_value value
+    end
+    
+    # Return a JSON 'object' for the various data values. Allows for
+    # persistence of the cookie information
+    #
+    # @param [Array] a options controlling output JSON text 
+    #   (usually a State and a depth)
+    # @return [String] JSON representation of object data
+    def to_json *a
+      result = {
+        :json_class => self.class.name,
+        :name => @name,
+        :value => @value,
+        :domain => @domain,
+        :path => @path,
+        :created_at => @created_at
+      }
+      {
+        :expiry => @expiry,
+        :secure => (true if @secure),
+        :http_only => (true if @http_only),
+        :version => (@version if version != 0),
+        :comment => @comment,
+        :comment_url => @comment_url,
+        :discard => (true if @discard),
+        :ports => @ports
+      }.each do |name, value|
+        result[name] = value if value
+      end
+      result.to_json(*a)
+    end
+    
+    # Given a Hash representation of a JSON document, create a local cookie
+    # from the included data.
+    #
+    # @param [Hash] o JSON object of array data
+    # @return [Cookie] cookie formed from JSON data
+    def self.json_create o
+      params = o.inject({}) do |hash, (key, value)|
+        hash[key.to_sym] = value
+        hash
+      end
+      params[:version] ||= 0
+      params[:created_at] = Time.parse params[:created_at]
+      if params[:expiry].is_a? String
+        params[:expires_at] = Time.parse params[:expiry]
+      else
+        params[:max_age] = params[:expiry]
+      end
+      params.delete :expiry
+
+      self.new params
+    end
+    
+    # Compute the cookie search domains for a given request URI
+    # This will be the effective host of the request uri, along with any
+    # possibly matching dot-prefixed domains
+    #
+    # @param request_uri [String, URI] address being requested
+    # @return [Array<String>] String domain matches
+    def self.compute_search_domains request_uri
+      CookieValidation.compute_search_domains request_uri
+    end
+  protected
+    # Call {from_set_cookie} to create a new Cookie instance
+    def initialize args
+      
+      @created_at, @name, @value, @domain, @path, @secure, 
+      @http_only, @version, @comment, @comment_url, @discard, @ports \
+      = args.values_at \
+      :created_at, :name, :value, :domain, :path, :secure, 
+      :http_only, :version, :comment, :comment_url, :discard, :ports
+
+      @created_at ||= Time.now
+      @expiry     = args[:max_age]   || args[:expires_at] 
+      @secure     ||= false
+      @http_only  ||= false
+      @discard    ||= false
+           
+      if @ports.is_a? Integer
+        @ports = [@ports]
+      end
+    end
+  end
+end

data/lib/cookiejar/cookie_validation.rb

@@ -0,0 +1,400 @@
+require 'cgi'
+require 'uri'
+module CookieJar
+  # Represents a set of cookie validation errors
+  class InvalidCookieError < StandardError
+    # [Array<String>] the specific validation issues encountered
+    attr_reader :messages
+    
+    # Create a new instance
+    # @param [String, Array<String>] the validation issue(s) encountered
+    def initialize message
+     if message.is_a? Array
+       @messages = message
+       message = message.join ', '
+     else
+       @messages = [message]
+     end
+     super message
+    end
+  end
+  
+  # Contains logic to parse and validate cookie headers
+  module CookieValidation
+    module PATTERN
+      include URI::REGEXP::PATTERN
+
+      TOKEN = '[^(),\/<>@;:\\\"\[\]?={}\s]+'
+      VALUE1 = "([^;]*)"
+      IPADDR = "#{IPV4ADDR}|#{IPV6ADDR}"
+      BASE_HOSTNAME = "(?:#{DOMLABEL}\\.)(?:((?:(?:#{DOMLABEL}\\.)+(?:#{TOPLABEL}\\.?))|local))"
+
+      QUOTED_PAIR = "\\\\[\\x00-\\x7F]"
+      LWS = "\\r\\n(?:[ \\t]+)"
+      # TEXT="[\\t\\x20-\\x7E\\x80-\\xFF]|(?:#{LWS})"
+      QDTEXT="[\\t\\x20-\\x21\\x23-\\x7E\\x80-\\xFF]|(?:#{LWS})"
+      QUOTED_TEXT = "\\\"(?:#{QDTEXT}|#{QUOTED_PAIR})*\\\""
+      VALUE2 = "#{TOKEN}|#{QUOTED_TEXT}"
+
+    end
+    BASE_HOSTNAME = /#{PATTERN::BASE_HOSTNAME}/
+    BASE_PATH = /\A((?:[^\/?#]*\/)*)/
+    IPADDR = /\A#{PATTERN::IPADDR}\Z/
+    HDN = /\A#{PATTERN::HOSTNAME}\Z/
+    TOKEN = /\A#{PATTERN::TOKEN}\Z/
+    PARAM1 = /\A(#{PATTERN::TOKEN})(?:=#{PATTERN::VALUE1})?\Z/
+    PARAM2 = Regexp.new "(#{PATTERN::TOKEN})(?:=(#{PATTERN::VALUE2}))?(?:\\Z|;)", '', 'n'
+    # TWO_DOT_DOMAINS = /\A\.(com|edu|net|mil|gov|int|org)\Z/
+    
+    # Converts the input object to a URI (if not already a URI)
+    #
+    # @param [String, URI] request_uri URI we are normalizing
+    # @param [URI] URI representation of input string, or original URI
+    def self.to_uri request_uri
+      (request_uri.is_a? URI)? request_uri : (URI.parse request_uri)
+    end
+    
+    # Converts an input cookie or uri to a string representing the path.
+    # Assume strings are already paths
+    #
+    # @param [String, URI, Cookie] object containing the path
+    # @return [String] path information
+    def self.to_path uri_or_path
+      if (uri_or_path.is_a? URI) || (uri_or_path.is_a? Cookie)
+        uri_or_path.path
+      else
+        uri_or_path
+      end
+    end
+    
+    # Converts an input cookie or uri to a string representing the domain.
+    # Assume strings are already domains. Value may not be an effective host.
+    # 
+    # @param [String, URI, Cookie] object containing the domain
+    # @return [String] domain information.
+    def self.to_domain uri_or_domain
+      if uri_or_domain.is_a? URI
+        uri_or_domain.host
+      elsif uri_or_domain.is_a? Cookie
+        uri_or_domain.domain
+      else
+        uri_or_domain
+      end
+    end
+    
+    # Compare a tested domain against the base domain to see if they match, or
+    # if the base domain is reachable.
+    #
+    # @param [String] tested_domain domain to be tested against
+    # @param [String] base_domain new domain being tested
+    # @return [String,nil] matching domain on success, nil on failure
+    def self.domains_match tested_domain, base_domain
+      base = effective_host base_domain
+      search_domains = compute_search_domains_for_host base
+      result = search_domains.find do |domain| 
+        domain == tested_domain  
+      end
+    end
+    
+    # Compute the reach of a hostname (RFC 2965, section 1)
+    # Determines the next highest superdomain
+    #
+    # @param [String,URI,Cookie] hostname hostname, or object holding hostname
+    # @return [String,nil] next highest hostname, or nil if none
+    def self.hostname_reach hostname
+      host = to_domain hostname
+      host = host.downcase
+      match = BASE_HOSTNAME.match host
+      if match
+        match[1]
+      end
+    end
+        
+    # Compute the base of a path, for default cookie path assignment
+    #
+    # @param [String, URI, Cookie] path, or object holding path
+    # @return base path (all characters up to final '/')
+    def self.cookie_base_path path
+      BASE_PATH.match(to_path path)[1]
+    end
+    
+    # Processes cookie path data using the following rules:
+    # Paths are separated by '/' characters, and accepted values are truncated
+    # to the last '/' character. If no path is specified in the cookie, a path
+    # value will be taken from the request URI which was used for the site.
+    #
+    # Note that this will not attempt to detect a mismatch of the request uri domain
+    # and explicitly specified cookie path
+    #
+    # @param [String,URI] request URI yielding this cookie
+    # @param [String] path on cookie
+    def self.determine_cookie_path request_uri, cookie_path
+      uri = to_uri request_uri
+      cookie_path = to_path cookie_path
+      
+      if cookie_path == nil || cookie_path.empty?
+        cookie_path = cookie_base_path uri.path
+      end
+      cookie_path
+    end
+    
+    # Given a URI, compute the relevant search domains for pre-existing
+    # cookies. This includes all the valid dotted forms for a named or IP
+    # domains.
+    # 
+    # @param [String, URI] request_uri requested uri
+    # @return [Array<String>] all cookie domain values which would match the
+    #   requested uri
+    def self.compute_search_domains request_uri
+      uri = to_uri request_uri
+      host = uri.host
+      compute_search_domains_for_host host
+    end
+    
+    # Given a host, compute the relevant search domains for pre-existing
+    # cookies
+    #
+    # @param [String] host host being requested
+    # @return [Array<String>] all cookie domain values which would match the
+    #   requested uri
+    def self.compute_search_domains_for_host host
+      host = effective_host host
+      result = [host]
+      unless host =~ IPADDR
+        result << ".#{host}"
+        base = hostname_reach host
+        if base
+          result << ".#{base}"
+        end
+      end
+      result
+    end
+    
+    # Processes cookie domain data using the following rules:
+    # Domains strings of the form .foo.com match 'foo.com' and all immediate
+    # subdomains of 'foo.com'. Domain strings specified of the form 'foo.com' are
+    # modified to '.foo.com', and as such will still apply to subdomains.
+    #
+    # Cookies without an explicit domain will have their domain value taken directly
+    # from the URL, and will _NOT_ have any leading dot applied. For example, a request
+    # to http://foo.com/ will cause an entry for 'foo.com' to be created - which applies
+    # to foo.com but no subdomain.
+    #
+    # Note that this will not attempt to detect a mismatch of the request uri domain
+    # and explicitly specified cookie domain
+    #
+    # @param [String, URI] request_uri originally requested URI
+    # @param [String] cookie domain value
+    # @return [String] effective host
+    def self.determine_cookie_domain request_uri, cookie_domain
+      uri = to_uri request_uri
+      domain = to_domain cookie_domain
+    
+      if domain == nil || domain.empty?
+        domain = effective_host uri.host
+      else
+        domain = domain.downcase
+        if domain =~ IPADDR || domain.start_with?('.')
+          domain
+        else
+          ".#{domain}" 
+        end
+      end
+    end
+    
+    # Compute the effective host (RFC 2965, section 1)
+    #
+    # Has the added additional logic of searching for interior dots specifically, and
+    # matches colons to prevent .local being suffixed on IPv6 addresses
+    #
+    # @param [String, URI] host_or_uridomain name, or absolute URI
+    # @return [String] effective host per RFC rules
+    def self.effective_host host_or_uri
+      hostname = to_domain host_or_uri
+      hostname = hostname.downcase
+    
+      if /.[\.:]./.match(hostname) || hostname == '.local'
+        hostname
+      else
+        hostname + '.local'
+      end
+    end
+    
+    # Check whether a cookie meets all of the rules to be created, based on 
+    # its internal settings and the URI it came from.
+    #
+    # @param [String,URI] request_uri originally requested URI
+    # @param [Cookie] cookie object
+    # @param [true] will always return true on success
+    # @raise [InvalidCookieError] on failures, containing all validation errors
+    def self.validate_cookie request_uri, cookie
+      uri = to_uri request_uri
+      request_host = effective_host uri.host
+      request_path = uri.path
+      request_secure = (uri.scheme == 'https')
+      cookie_host = cookie.domain
+      cookie_path = cookie.path
+      
+      errors = []
+    
+      # From RFC 2965, Section 3.3.2 Rejecting Cookies
+    
+      # A user agent rejects (SHALL NOT store its information) if the 
+      # Version attribute is missing. Note that the legacy Set-Cookie
+      # directive will result in an implicit version 0.
+      unless cookie.version
+        errors << "Version missing"
+      end
+
+      # The value for the Path attribute is not a prefix of the request-URI
+      unless request_path.start_with? cookie_path 
+        errors << "Path is not a prefix of the request uri path"
+      end
+
+      unless cookie_host =~ IPADDR || #is an IPv4 or IPv6 address
+        cookie_host =~ /.\../ || #contains an embedded dot
+        cookie_host == '.local' #is the domain cookie for local addresses
+        errors << "Domain format is illegal"
+      end
+    
+      # The effective host name that derives from the request-host does
+      # not domain-match the Domain attribute.
+      #
+      # The request-host is a HDN (not IP address) and has the form HD,
+      # where D is the value of the Domain attribute, and H is a string
+      # that contains one or more dots.
+      unless domains_match cookie_host, uri
+        errors << "Domain is inappropriate based on request URI hostname"
+      end
+    
+      # The Port attribute has a "port-list", and the request-port was
+      # not in the list.
+      unless cookie.ports.nil? || cookie.ports.length != 0
+        unless cookie.ports.find_index uri.port
+          errors << "Ports list does not contain request URI port"
+        end
+      end
+
+      raise (InvalidCookieError.new errors) unless errors.empty?
+
+      # Note: 'secure' is not explicitly defined as an SSL channel, and no
+      # test is defined around validity and the 'secure' attribute
+      true
+    end
+    
+    # Break apart a traditional (non RFC 2965) cookie value into its core
+    # components. This does not do any validation, or defaulting of values
+    # based on requested URI
+    # 
+    # @param [String] set_cookie_value a Set-Cookie header formatted cookie
+    #   definition
+    # @return [Hash] Contains the parsed values of the cookie
+    def self.parse_set_cookie set_cookie_value
+      args = { }
+      params=set_cookie_value.split /;\s*/
+      
+      first=true
+      params.each do |param|
+        result = PARAM1.match param
+        if !result
+          raise InvalidCookieError.new "Invalid cookie parameter in cookie '#{set_cookie_value}'"
+        end
+        key = result[1].downcase.to_sym
+        keyvalue = result[2]
+        if first
+          args[:name] = result[1]
+          args[:value] = keyvalue
+          first = false
+        else
+          case key
+          when :expires
+            args[:expires_at] = Time.parse keyvalue
+          when *[:domain, :path]
+            args[key] = keyvalue
+          when :secure
+            args[:secure] = true
+          when :httponly
+            args[:http_only] = true
+          else
+            raise InvalidCookieError.new "Unknown cookie parameter '#{key}'"
+          end
+        end
+      end
+      args[:version] = 0
+      args
+    end
+    
+    # Parse a RFC 2965 value and convert to a literal string
+    def self.value_to_string value
+      if /\A"(.*)"\Z/.match value
+        value = $1
+        value = value.gsub(/\\(.)/, '\1')
+      else
+        value
+      end
+    end
+    
+    # Attempt to decipher a partially decoded version of text cookie values
+    def self.decode_value value
+      if /\A"(.*)"\Z/.match value
+        value_to_string value
+      else
+        CGI.unescape value
+      end
+    end
+      
+    # Break apart a RFC 2965 cookie value into its core components. 
+    # This does not do any validation, or defaulting of values
+    # based on requested URI
+    # 
+    # @param [String] set_cookie_value a Set-Cookie2 header formatted cookie
+    #   definition
+    # @return [Hash] Contains the parsed values of the cookie
+    def self.parse_set_cookie2 set_cookie_value
+      args = { }
+      first = true
+      index = 0
+      begin
+        md = PARAM2.match set_cookie_value, index
+        if md.nil? || md.offset(0).first != index
+          raise InvalidCookieError.new "Invalid Set-Cookie2 header '#{set_cookie_value}'"
+        end
+        index=md.offset(0)[1]
+        
+        key = md[1].downcase.to_sym
+        keyvalue = md[2] || md[3]
+        if first
+          args[:name] = md[1]
+          args[:value] = keyvalue
+          first = false
+        else
+          keyvalue = value_to_string keyvalue
+          case key
+          when *[:comment,:commenturl,:domain,:path]
+            args[key] = keyvalue
+          when *[:discard,:secure]
+            args[key] = true
+          when :httponly
+            args[:http_only] = true
+          when :"max-age"
+            args[:max_age] = keyvalue.to_i
+          when :version
+            args[:version] = keyvalue.to_i
+          when :port
+            # must be in format '"port,port"'
+            ports = keyvalue.split /,\s*/
+            args[:ports] = ports.map do |portstr| portstr.to_i end
+          else
+            raise InvalidCookieError.new "Unknown cookie parameter '#{key}'"
+          end
+        end
+      end until md.post_match.empty?
+      # if our last match in the scan failed
+      if args[:version] != 1
+        raise InvalidCookieError.new "Set-Cookie2 declares a non RFC2965 version cookie"
+      end
+      
+      args
+    end    
+  end
+end