Fixes a bug that caused relative links not to be properly absolutifie…

…d in the presence of a <base> tag. Fixes #46
jaimeiniesta · Oct 1, 2013 · 571a13b · 571a13b
1 parent c0ef38c
commit 571a13b
Show file tree

Hide file tree

Showing 4 changed files with 50 additions and 1 deletion.
diff --git a/lib/meta_inspector/scraper.rb b/lib/meta_inspector/scraper.rb
@@ -234,12 +234,23 @@ def absolutify_url(uri)
       if uri =~ /^\w*\:/i
         normalize_url(uri)
       else
-        Addressable::URI.join(@url, uri).normalize.to_s
+        Addressable::URI.join(base_url, uri).normalize.to_s
       end
     rescue URI::InvalidURIError, Addressable::URI::InvalidURIError => e
       add_fatal_error "Link parsing exception: #{e.message}" and nil
     end
 
+    # Returns the base url to absolutify relative links. This can be the one set on a <base> tag,
+    # or the url of the document if no <base> tag was found.
+    def base_url
+      base_href || @url
+    end
+
+    # Returns the value of the href attribute on the <base /> tag, if it exists
+    def base_href
+      parsed_document.search('base').first.attributes['href'].value rescue nil
+    end
+
     # Convert a protocol-relative url to its full form, depending on the scheme of the page that contains it
     def unrelativize_url(url)
       url =~ /^\/\// ? "#{scheme}://#{url[2..-1]}" : url

diff --git a/spec/fixtures/relative_links_with_base.response b/spec/fixtures/relative_links_with_base.response
@@ -0,0 +1,22 @@
+HTTP/1.1 200 OK
+Server: nginx/1.0.5
+Date: Thu, 29 Dec 2011 23:10:13 GMT
+Content-Type: text/html
+Content-Length: 15013
+Last-Modified: Fri, 02 Dec 2011 21:00:49 GMT
+Connection: keep-alive
+Accept-Ranges: bytes
+
+<!DOCTYPE html>
+<html>
+<head>
+  <base href="http://relativewithbase.com/" />
+  <meta charset="utf-8" />
+  <title>Relative links</title>
+</head>
+<body>
+  <p>Relative links</p>
+  <a href="about">About</a>
+  <a href="../sitemap">Sitemap</a>
+</body>
+</html>
diff --git a/spec/metainspector_spec.rb b/spec/metainspector_spec.rb
@@ -272,6 +272,18 @@
     end
   end
 
+  describe 'Relative links with base' do
+    it 'should get the relative links from a document' do
+      m = MetaInspector.new('http://relativewithbase.com/company/page2')
+      m.internal_links.should == ['http://relativewithbase.com/about', 'http://relativewithbase.com/sitemap']
+    end
+
+    it 'should get the relative links from a directory' do
+      m = MetaInspector.new('http://relativewithbase.com/company/page2/')
+      m.internal_links.should == ['http://relativewithbase.com/about', 'http://relativewithbase.com/sitemap']
+    end
+  end
+
   describe 'Non-HTTP links' do
     before(:each) do
       @m = MetaInspector.new('http://example.com/nonhttp')

diff --git a/spec/spec_helper.rb b/spec/spec_helper.rb
@@ -47,6 +47,10 @@ def fixture_file(filename)
 FakeWeb.register_uri(:get, "http://relative.com/company", :response => fixture_file("relative_links.response"))
 FakeWeb.register_uri(:get, "http://relative.com/company/", :response => fixture_file("relative_links.response"))
 
+FakeWeb.register_uri(:get, "http://relativewithbase.com/",                :response => fixture_file("relative_links_with_base.response"))
+FakeWeb.register_uri(:get, "http://relativewithbase.com/company/page2",   :response => fixture_file("relative_links_with_base.response"))
+FakeWeb.register_uri(:get, "http://relativewithbase.com/company/page2/",  :response => fixture_file("relative_links_with_base.response"))
+
 # These examples are used to test the redirections from HTTP to HTTPS and vice versa
 # http://facebook.com => https://facebook.com
 FakeWeb.register_uri(:get, "http://facebook.com/",          :response => fixture_file("facebook.com.response"))