RubyGems - metainspector - Versions diffs - 1.13.0 → 1.13.1 - Mend

metainspector 1.13.0 → 1.13.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (6) hide show

data/lib/meta_inspector/scraper.rb +9 -2
data/lib/meta_inspector/version.rb +1 -1
data/spec/fixtures/malformed_href.response +27 -0
data/spec/metainspector_spec.rb +36 -14
data/spec/spec_helper.rb +1 -0
metadata +5 -4

data/lib/meta_inspector/scraper.rb CHANGED Viewed

@@ -53,12 +53,12 @@ module MetaInspector
     # Internal links found on the page, as absolute URLs
     def internal_links
-      @internal_links ||= links.select {|link| URI.parse(link).host == host }
+      @internal_links ||= links.select {|link| host_from_url(link) == host }
     end
     # External links found on the page, as absolute URLs
     def external_links
-      @external_links ||= links.select {|link| URI.parse(link).host != host }
+      @external_links ||= links.select {|link| host_from_url(link) != host }
     end
     # Images found on the page, as absolute URLs
@@ -245,6 +245,13 @@ module MetaInspector
       url =~ /^\/\// ? "#{scheme}://#{url[2..-1]}" : url
     end
+    # Extracts the host from a given URL
+    def host_from_url(url)
+      URI.parse(url).host
+    rescue URI::InvalidURIError, URI::InvalidComponentError => e
+      add_fatal_error "Link parsing exception: #{e.message}" and nil
+    end
     # Look for the first <p> block with 120 characters or more
     def secondary_description
       first_long_paragraph = parsed_document.search('//p[string-length() >= 120]').first

data/lib/meta_inspector/version.rb CHANGED Viewed

@@ -1,5 +1,5 @@
 # -*- encoding: utf-8 -*-
 module MetaInspector
-  VERSION = "1.13.0"
+  VERSION = "1.13.1"
 end

data/spec/fixtures/malformed_href.response ADDED Viewed

@@ -0,0 +1,27 @@
+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>
+    <title>Malformed hrefs</title>
+  </head>
+  <body>
+    <h1>Good links</h1>
+    <a href="/faqs">FAQs</a>
+    <a href="skype:joeuser?call">a skype link</a>
+    <a href="telnet://telnet.cdrom.com">a telnet link</a>
+    <a href="javascript:alert('ok');">ok</a>
+    <h1>Bad links due to malformed href</h1>
+    <a href="javascript://">oops</a>
+    <a href="mailto:email(at)example.com">
+  </body>
+</html>

data/spec/metainspector_spec.rb CHANGED Viewed

@@ -177,22 +177,44 @@ describe MetaInspector do
                             "http://example.com/search?q=espa%C3%B1a#top"]
       end
-      it "should get correct internal links, encoding the URLs as needed but respecting # and ?" do
-        m = MetaInspector.new('http://international.com')
-        m.internal_links.should == [ "http://international.com/espa%C3%B1a.asp",
-                                     "http://international.com/roman%C3%A9e",
-                                     "http://international.com/faqs#cami%C3%B3n",
-                                     "http://international.com/search?q=cami%C3%B3n",
-                                     "http://international.com/search?q=espa%C3%B1a#top"]
+      describe "internal links" do
+        it "should get correct internal links, encoding the URLs as needed but respecting # and ?" do
+          m = MetaInspector.new('http://international.com')
+          m.internal_links.should == [ "http://international.com/espa%C3%B1a.asp",
+                                       "http://international.com/roman%C3%A9e",
+                                       "http://international.com/faqs#cami%C3%B3n",
+                                       "http://international.com/search?q=cami%C3%B3n",
+                                       "http://international.com/search?q=espa%C3%B1a#top"]
+        end
+        it "should not crash when processing malformed hrefs" do
+          m = MetaInspector.new('http://example.com/malformed_href')
+          expect {
+            m.internal_links.should == [ "http://example.com/faqs" ]
+            m.should_not be_ok
+          }.to_not raise_error
+        end
       end
-      it "should get correct external links, encoding the URLs as needed but respecting # and ?" do
-        m = MetaInspector.new('http://international.com')
-        m.external_links.should == [ "http://example.com/espa%C3%B1a.asp",
-                                     "http://example.com/roman%C3%A9e",
-                                     "http://example.com/faqs#cami%C3%B3n",
-                                     "http://example.com/search?q=cami%C3%B3n",
-                                     "http://example.com/search?q=espa%C3%B1a#top"]
+      describe "external links" do
+        it "should get correct external links, encoding the URLs as needed but respecting # and ?" do
+          m = MetaInspector.new('http://international.com')
+          m.external_links.should == [ "http://example.com/espa%C3%B1a.asp",
+                                       "http://example.com/roman%C3%A9e",
+                                       "http://example.com/faqs#cami%C3%B3n",
+                                       "http://example.com/search?q=cami%C3%B3n",
+                                       "http://example.com/search?q=espa%C3%B1a#top"]
+        end
+        it "should not crash when processing malformed hrefs" do
+          m = MetaInspector.new('http://example.com/malformed_href')
+          expect {
+            m.external_links.should == ["skype:joeuser?call", "telnet://telnet.cdrom.com",
+                                        "javascript:alert('ok');", "javascript://",
+                                        "mailto:email(at)example.com"]
+            m.should_not be_ok
+          }.to_not raise_error
+        end
       end
     end

data/spec/spec_helper.rb CHANGED Viewed

@@ -29,6 +29,7 @@ FakeWeb.register_uri(:get, "http://protocol-relative.com", :response => fixture_
 FakeWeb.register_uri(:get, "https://protocol-relative.com", :response => fixture_file("protocol_relative.response"))
 FakeWeb.register_uri(:get, "http://example.com/nonhttp", :response => fixture_file("nonhttp.response"))
 FakeWeb.register_uri(:get, "http://example.com/invalid_href", :response => fixture_file("invalid_href.response"))
+FakeWeb.register_uri(:get, "http://example.com/malformed_href", :response => fixture_file("malformed_href.response"))
 FakeWeb.register_uri(:get, "http://www.youtube.com/watch?v=iaGSSrp49uc", :response => fixture_file("youtube.response"))
 FakeWeb.register_uri(:get, "http://markupvalidator.com/faqs", :response => fixture_file("markupvalidator_faqs.response"))
 FakeWeb.register_uri(:get, "https://twitter.com/markupvalidator", :response => fixture_file("twitter_markupvalidator.response"))

metadata CHANGED Viewed

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: metainspector
 version: !ruby/object:Gem::Version
-  hash: 35
+  hash: 33
   prerelease:
   segments:
   - 1
   - 13
-  - 0
-  version: 1.13.0
+  - 1
+  version: 1.13.1
 platform: ruby
 authors:
 - Jaime Iniesta
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
-date: 2012-12-03 00:00:00 Z
+date: 2012-12-13 00:00:00 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -164,6 +164,7 @@ files:
 - spec/fixtures/international.response
 - spec/fixtures/invalid_href.response
 - spec/fixtures/iteh.at.response
+- spec/fixtures/malformed_href.response
 - spec/fixtures/markupvalidator_faqs.response
 - spec/fixtures/nonhttp.response
 - spec/fixtures/pagerankalert.com.response