Fix bug where metainspector would crash if no links found on the page

jaimeiniesta · Aug 12, 2012 · 80c68ad · 80c68ad
1 parent 03ead0d
commit 80c68ad
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 1 deletion.
diff --git a/lib/meta_inspector/scraper.rb b/lib/meta_inspector/scraper.rb
@@ -149,7 +149,7 @@ def method_missing(method_name)
     def parsed_links
       @parsed_links ||= parsed_document.search("//a") \
                         .map {|link| link.attributes["href"] \
-                        .to_s.strip}.uniq rescue nil
+                        .to_s.strip}.uniq rescue []
     end
 
     def parsed_images

diff --git a/spec/fixtures/empty_page.response b/spec/fixtures/empty_page.response
@@ -0,0 +1,17 @@
+HTTP/1.1 200 OK
+Server: nginx/0.7.67
+Date: Fri, 18 Nov 2011 21:46:46 GMT
+Content-Type: text/html
+Connection: keep-alive
+Last-Modified: Mon, 14 Nov 2011 16:53:18 GMT
+Content-Length: 4987
+X-Varnish: 2000423390
+Age: 0
+Via: 1.1 varnish
+
+<html>
+  <head>
+  </head>
+  <body>
+  </body>
+</html>
diff --git a/spec/metainspector_spec.rb b/spec/metainspector_spec.rb
@@ -16,6 +16,7 @@
   FakeWeb.register_uri(:get, "http://www.youtube.com/watch?v=iaGSSrp49uc", :response => fixture_file("youtube.response"))
   FakeWeb.register_uri(:get, "http://w3clove.com/faqs", :response => fixture_file("w3clove_faqs.response"))
   FakeWeb.register_uri(:get, "https://twitter.com/w3clove", :response => fixture_file("twitter_w3clove.response"))
+  FakeWeb.register_uri(:get, "https://example.com/empty", :response => fixture_file("empty_page.response"))
 
   describe 'Initialization' do
     it 'should accept an URL with a scheme' do
@@ -164,6 +165,11 @@
                           "http://twitter.com/W3CLove",
                           "http://us4.campaign-archive1.com/home/?u=6af3ab69c286561d0f0f25671&id=04a0dab609" ]
     end
+
+    it "should return empty array if no links found" do
+      m = MetaInspector.new('http://example.com/empty')
+      m.links.should == []
+    end
   end
 
   describe 'Non-HTTP links' do