postrank-uri 1.0.13 → 1.0.14

data/lib/postrank-uri.rb

@@ -18,6 +18,22 @@ module Addressable
       dom = dp.public_suffix
       dom = dp.domain.downcase + "." + dom unless dp.domain.empty?
     end
+
+    def normalized_query
+      @normalized_query ||= (begin
+        if self.query && self.query.strip != ''
+          (self.query.strip.split("&", -1).map do |pair|
+            Addressable::URI.normalize_component(
+              pair,
+              Addressable::URI::CharacterClasses::QUERY.sub("\\&", "")
+            )
+          end).join("&")
+        else
+          nil
+        end
+      end)
+    end
+
   end
 end
 
@@ -33,7 +49,7 @@ module PostRank
     URIREGEX = {}
     URIREGEX[:protocol] = /https?:\/\//i
     URIREGEX[:valid_preceding_chars] = /(?:|\.|[^-\/"':!=A-Z0-9_@＠]|^|\:)/i
-    URIREGEX[:valid_domain] = /(?:[^[:punct:]\s][\.-](?=[^[:punct:]\s])|[^[:punct:]\s]){1,}\.[a-z]{2,}(?::[0-9]+)?/i
+    URIREGEX[:valid_domain] = /\b(?:[a-z0-9-]{1,63}\.){1,}[a-z]{2,63}(?::[0-9]+)?/i
     URIREGEX[:valid_general_url_path_chars] = /[a-z0-9!\*';:=\+\,\$\/%#\[\]\-_~]/i
 
     # Allow URL paths to contain balanced parens
@@ -203,8 +219,7 @@ module PostRank
       end
 
       uri.scheme = 'http' if uri.host && !uri.scheme
-
-      uri.normalize
+      uri.normalize!
     end
 
   end

data/lib/postrank-uri/version.rb

@@ -1,5 +1,5 @@
 module PostRank
   module URI
-    VERSION = "1.0.13"
+    VERSION = "1.0.14"
   end
 end

data/postrank-uri.gemspec

@@ -18,6 +18,7 @@ Gem::Specification.new do |s|
   s.add_dependency "domainatrix"
   s.add_dependency "nokogiri"
   s.add_development_dependency "rspec"
+  #s.add_development_dependency "idn" # test with idn
 
   s.files         = `git ls-files`.split("\n")
   s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")

data/spec/postrank-uri_spec.rb

@@ -243,6 +243,11 @@ describe PostRank::URI do
       u.should include('http://abc.co/')
     end
 
+    it "should pickup urls inside tags" do
+      u = e("<a href='http://bit.ly/3fds3'>abc.com</a>")
+      u.should include('http://abc.com/')
+    end
+
     context "multibyte characters" do
       it "should stop extracting URLs at the full-width CJK space character" do
         e("http://www.youtube.com/watch?v=w_j4Lda25jA　　とんかつ定食").should == ["http://www.youtube.com/watch?v=w_j4Lda25jA"]
@@ -298,7 +303,8 @@ describe PostRank::URI do
         "http://alex.pages.examplecom" => nil,
         "example" => nil,
         "http://127.0.0.1" => nil,
-        "localhost" => nil
+        "localhost" => nil,
+        "hello-there.com/you" => "hello-there.com"
       }
 
       url_list.each_pair do |url, expected_result|
@@ -309,4 +315,10 @@ describe PostRank::URI do
       end
     end
   end
+
+  context "parse" do
+    it 'should not fail on large host-part look-alikes' do
+      PostRank::URI.parse('a'*64+'.ca').host.should == nil
+    end
+  end
 end

metadata

@@ -2,7 +2,7 @@
 name: postrank-uri
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 1.0.13
+  version: 1.0.14
 platform: ruby
 authors: 
 - Ilya Grigorik
@@ -10,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-05-10 00:00:00 -04:00
+date: 2011-05-18 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency