postrank-uri 1.0.7 → 1.0.8

data/lib/postrank-uri/c18n.yml

@@ -11,9 +11,11 @@
 :hosts:
   nytimes.com:
   - partner
+  - pagewanted
   - emc
   - _r
   - ref
+  - src
   diepresse.com:
   - _vl_backlink
   washingtonpost.com:
@@ -41,4 +43,8 @@
   cnet.com:
   - part
   - subj
-  - tag
+  - tag
+  wsj.com:
+  - mod
+  allthingsd.com:
+  - mod

data/lib/postrank-uri/version.rb

@@ -1,5 +1,5 @@
 module PostRank
   module URI
-    VERSION = "1.0.7"
+    VERSION = "1.0.8"
   end
 end

data/lib/postrank-uri.rb

@@ -125,6 +125,7 @@ module PostRank
     def normalize(uri)
       u = parse(uri)
       u.path = u.path.squeeze('/')
+      u.path = u.path.chomp('/') if u.path.size != 1
       u.query = nil if u.query && u.query.empty?
       u.fragment = nil
       u

data/spec/c18n_hosts.yml

@@ -5,11 +5,23 @@
 - - http://www.nytimes.com/2011/02/20/magazine/20FOB-Medium-t.html?ref=magazine
   - http://www.nytimes.com/2011/02/20/magazine/20FOB-Medium-t.html
 
+- - http://www.nytimes.com/2011/03/13/business/13hire.html?pagewanted=1&_r=1&ref=technology
+  - http://www.nytimes.com/2011/03/13/business/13hire.html
+
+- - http://www.nytimes.com/2011/03/15/business/media/15adco.html?_r=2&src=recg
+  - http://www.nytimes.com/2011/03/15/business/media/15adco.html
+
+- - http://networkeffect.allthingsd.com/20110308/googles-approach-to-social/?mod=tweet
+  - http://networkeffect.allthingsd.com/20110308/googles-approach-to-social
+
+- - http://online.wsj.com/article/SB10001424052748704657704576150191661959856.html?mod=WSJ_hp_LEFTWhatsNewsCollection
+  - http://online.wsj.com/article/SB10001424052748704657704576150191661959856.html
+
 - - http://diepresse.com/home/wirtschaft/636448/Griechenland_Drachme-als-letzte-Rettung?_vl_backlink=%2Fhome
   - http://diepresse.com/home/wirtschaft/636448/Griechenland_Drachme-als-letzte-Rettung
 
 - - http://dotearth.blogs.nytimes.com/2010/12/14/beyond-political-science/?partner=rss&emc=rss
-  - http://dotearth.blogs.nytimes.com/2010/12/14/beyond-political-science/
+  - http://dotearth.blogs.nytimes.com/2010/12/14/beyond-political-science
 
 - - http://www.washingtonpost.com/wp-dyn/content/article/2010/12/14/AR2010121406045.html?nav=rss_email/components
   - http://www.washingtonpost.com/wp-dyn/content/article/2010/12/14/AR2010121406045.html
@@ -36,7 +48,7 @@
   - http://www.dw-world.de/dw/article/0,,6330472,00.html
 
 - - http://www.repubblica.it/rubriche/il-caso-del-giorno/2010/12/13/news/riscossa_aeffe-10153565/?rss
-  - http://www.repubblica.it/rubriche/il-caso-del-giorno/2010/12/13/news/riscossa_aeffe-10153565/
+  - http://www.repubblica.it/rubriche/il-caso-del-giorno/2010/12/13/news/riscossa_aeffe-10153565
 
 - - http://www.welt.de/sport/Der-Hoellenritt-des-Fussball-Profis-Jean-Marc-Bosman.html?wtmc=RSS.Sport.Fussball
   - http://www.welt.de/sport/Der-Hoellenritt-des-Fussball-Profis-Jean-Marc-Bosman.html

data/spec/postrank-uri_spec.rb

@@ -86,6 +86,16 @@ describe PostRank::URI do
       n('IGVITA.COM/ABC').should == (igvita + "ABC")
     end
 
+    it "should remove trailing slash on paths" do
+      n('http://igvita.com/').should == 'http://igvita.com/'
+
+      n('http://igvita.com/a').should == 'http://igvita.com/a'
+      n('http://igvita.com/a/').should == 'http://igvita.com/a'
+
+      n('http://igvita.com/a/b').should == 'http://igvita.com/a/b'
+      n('http://igvita.com/a/b/').should == 'http://igvita.com/a/b'
+    end
+
   end
 
   context "canonicalization" do
@@ -125,7 +135,7 @@ describe PostRank::URI do
     context "embedded links" do
       it "should extract embedded redirects from Google News" do
         u = c('http://news.google.com/news/url?sa=t&fd=R&&url=http://www.ctv.ca/CTVNews/Politics/20110111/')
-        u.should == 'http://www.ctv.ca/CTVNews/Politics/20110111/'
+        u.should == 'http://www.ctv.ca/CTVNews/Politics/20110111'
       end
 
       it "should extract embedded redirects from xfruits.com" do
@@ -135,7 +145,7 @@ describe PostRank::URI do
 
       it "should extract embedded redirects from MySpace" do
         u = c('http://www.myspace.com/Modules/PostTo/Pages/?u=http%3A%2F%2Fghanaian-chronicle.com%2Fnews%2Fother-news%2Fcanadian-high-commissioner-urges-media%2F&t=Canadian%20High%20Commissioner%20urges%20media')
-        u.should == 'http://ghanaian-chronicle.com/news/other-news/canadian-high-commissioner-urges-media/'
+        u.should == 'http://ghanaian-chronicle.com/news/other-news/canadian-high-commissioner-urges-media'
       end
     end
   end
@@ -172,11 +182,12 @@ describe PostRank::URI do
     end
 
     it "should compute MD5 hash of the normalized URI" do
-      hash = '021a1032b1ea631a7c33d1a0ccc562bf'
+      hash = '55fae8910d312b7878a3201ed653b881'
 
-      h('http://EverBurnign.Com/feed/post/1').should == hash
-      h('Everburnign.com/feed/post/1').should == hash
-      h('everburnign.com/feed/post/1').should == hash
+      h('http://EverBurning.Com/feed/post/1').should == hash
+      h('Everburning.com/feed/post/1').should == hash
+      h('everburning.com/feed/post/1').should == hash
+      h('everburning.com/feed/post/1/').should == hash
     end
   end
 

metadata

@@ -1,12 +1,8 @@
 --- !ruby/object:Gem::Specification 
 name: postrank-uri
 version: !ruby/object:Gem::Version 
-  prerelease: false
-  segments: 
-  - 1
-  - 0
-  - 7
-  version: 1.0.7
+  prerelease: 
+  version: 1.0.8
 platform: ruby
 authors: 
 - Ilya Grigorik
@@ -14,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-02-22 00:00:00 -05:00
+date: 2011-03-31 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -25,10 +21,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 2
-        - 2
-        - 3
         version: 2.2.3
   type: :runtime
   version_requirements: *id001
@@ -40,8 +32,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 0
         version: "0"
   type: :runtime
   version_requirements: *id002
@@ -53,8 +43,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 0
         version: "0"
   type: :runtime
   version_requirements: *id003
@@ -66,8 +54,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 0
         version: "0"
   type: :development
   version_requirements: *id004
@@ -107,21 +93,17 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      segments: 
-      - 0
       version: "0"
 required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      segments: 
-      - 0
       version: "0"
 requirements: []
 
 rubyforge_project: postrank-uri
-rubygems_version: 1.3.7
+rubygems_version: 1.6.2
 signing_key: 
 specification_version: 3
 summary: URI normalization, c18n, escaping, and extraction