twingly-url 3.0.2 → 4.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 8c5e6153a97d4637dce5f7739d0297b760ca41f0
-  data.tar.gz: 4e05a760b5bb68f2cb0472a01d38e4f3f16cb713
+  metadata.gz: a0bef38a1195e09028dc139154dd6f7a833c9b91
+  data.tar.gz: 9abaafbd41a71c382f9833ddd6f9fcf5cd9ce08c
 SHA512:
-  metadata.gz: 410fd7c6c40896dbe6fc9450f89a97d65b1b53e263ed94e502bb986eb402716fb3366bc40c9b96f02d1458b19458971cd562e653e08345d24ac6c56f20751fd5
-  data.tar.gz: 94546494a6769b85a8c63fac77364b3a9075d7091198f4914ca065232e42c18310a0e409b2a224a54a199796b68a95cc525361d9839163f28591dbb6570b99b8
+  metadata.gz: b4e3602b1964b3b2b8f1a37b800f551bf316d8304258aeeb11a1ad43235c2ff81a265a640dbaae25540a9a732145af73543685a4e7f06b5115561d704f37d845
+  data.tar.gz: 44904c9aeae8aa885019c4fb646e3da8d6206d15ede5990df2371fdbebcf02ede04ae101557e6ee4b7b2d6f73f8f664643f7d0e9239fdb84f54ff520309af175

data/README.md

@@ -36,7 +36,7 @@ Run tests with
 
 You can get some profiling by running
 
-    bundle exec rake test:profile
+    bundle exec rake profile:normalize
 
 Note that this isn't a benchmark, we're using [ruby-prof] which will slow things down.
 

data/lib/twingly/url.rb

@@ -4,6 +4,7 @@ require "public_suffix"
 
 require_relative "url/null_url"
 require_relative "url/error"
+require_relative "version"
 
 PublicSuffix::List.private_domains = false
 
@@ -19,41 +20,66 @@ module Twingly
       IDN::Idna::IdnaError,
     ]
 
-    def self.parse(potential_url)
-      potential_url = String(potential_url)
-      potential_url = potential_url.scrub
-      potential_url = potential_url.strip
+    private_constant :ACCEPTED_SCHEMES, :ENDS_WITH_SLASH, :ERRORS
 
-      internal_parse(potential_url)
-    rescue Twingly::URL::Error, Twingly::URL::Error::ParseError => error
-      NullURL.new
-    end
+    class << self
+      def parse(potential_url)
+        internal_parse(potential_url)
+      rescue Twingly::URL::Error, Twingly::URL::Error::ParseError => error
+        NullURL.new
+      end
 
-    def self.internal_parse(potential_url)
-      addressable_uri = to_addressable_uri(potential_url)
-      raise Twingly::URL::Error::ParseError if addressable_uri.nil?
+      def internal_parse(potential_url)
+        addressable_uri = to_addressable_uri(potential_url)
+        raise Twingly::URL::Error::ParseError if addressable_uri.nil?
 
-      scheme = addressable_uri.scheme
-      raise Twingly::URL::Error::ParseError unless scheme =~ ACCEPTED_SCHEMES
+        scheme = addressable_uri.scheme
+        raise Twingly::URL::Error::ParseError unless scheme =~ ACCEPTED_SCHEMES
 
-      public_suffix_domain = PublicSuffix.parse(addressable_uri.display_uri.host)
-      raise Twingly::URL::Error::ParseError if public_suffix_domain.nil?
+        display_uri = addressable_display_uri(addressable_uri)
 
-      self.new(addressable_uri, public_suffix_domain)
-    rescue *ERRORS => error
-      error.extend(Twingly::URL::Error)
-      raise
-    end
+        public_suffix_domain = PublicSuffix.parse(display_uri.host)
+        raise Twingly::URL::Error::ParseError if public_suffix_domain.nil?
 
-    def initialize(addressable_uri, public_suffix_domain)
-      unless addressable_uri.is_a?(Addressable::URI)
-        raise ArgumentError, "First parameter must be an Addressable::URI"
+        raise Twingly::URL::Error::ParseError if public_suffix_domain.sld.nil?
+
+        new(addressable_uri, public_suffix_domain)
+      rescue *ERRORS => error
+        error.extend(Twingly::URL::Error)
+        raise
+      end
+
+      def to_addressable_uri(potential_url)
+       if potential_url.is_a?(Addressable::URI)
+          potential_url
+        else
+          potential_url = String(potential_url)
+          potential_url = potential_url.scrub
+          potential_url = potential_url.strip
+
+          Addressable::URI.heuristic_parse(potential_url)
+        end
       end
 
-      unless public_suffix_domain.is_a?(PublicSuffix::Domain)
-        raise ArgumentError, "Second parameter must be a PublicSuffix::Domain"
+      # Workaround for the following bug in addressable:
+      # https://github.com/sporkmonger/addressable/issues/224
+      def addressable_display_uri(addressable_uri)
+        addressable_uri.display_uri
+      rescue ArgumentError => error
+        if error.message.include?("invalid byte sequence in UTF-8")
+          raise Twingly::URL::Error::ParseError
+        end
+
+        raise
       end
 
+      private :new
+      private :internal_parse
+      private :to_addressable_uri
+      private :addressable_display_uri
+    end
+
+    def initialize(addressable_uri, public_suffix_domain)
       @addressable_uri      = addressable_uri
       @public_suffix_domain = public_suffix_domain
     end
@@ -63,7 +89,7 @@ module Twingly
     end
 
     def trd
-      public_suffix_domain.trd
+      public_suffix_domain.trd.to_s
     end
 
     def sld
@@ -101,7 +127,7 @@ module Twingly
       normalized_url.host   = normalized_host
       normalized_url.path   = normalized_path
 
-      self.class.internal_parse(normalized_url)
+      self.class.parse(normalized_url)
     end
 
     def normalized_scheme
@@ -143,15 +169,6 @@ module Twingly
       sprintf("#<%s:0x%x %s>", self.class.name, __id__, self.to_s)
     end
 
-    private_class_method \
-    def self.to_addressable_uri(potential_url)
-     if potential_url.is_a?(Addressable::URI)
-        potential_url
-      else
-        Addressable::URI.heuristic_parse(potential_url)
-      end
-    end
-
     private
 
     attr_reader :addressable_uri, :public_suffix_domain

data/lib/twingly/url/hasher.rb

@@ -1,5 +1,7 @@
 require 'digest'
 
+require_relative "../url"
+
 module Twingly
   class URL
     module Hasher

data/lib/twingly/version.rb

@@ -1,5 +1,5 @@
 module Twingly
   class URL
-    VERSION = "3.0.2"
+    VERSION = "4.0.0"
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: twingly-url
 version: !ruby/object:Gem::Version
-  version: 3.0.2
+  version: 4.0.0
 platform: ruby
 authors:
 - Twingly AB
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-11-11 00:00:00.000000000 Z
+date: 2016-02-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: addressable