Add support for base HTML element

Add support in the Page's to_absolute method for the base HTML element. This way it can correctly convert relative links for a given page document.
efrat-safanov · Jan 20, 2012 · 217a22b · 217a22b
1 parent 31a1515
commit 217a22b
Show file tree

Hide file tree

Showing 3 changed files with 69 additions and 2 deletions.
diff --git a/lib/anemone/page.rb b/lib/anemone/page.rb
@@ -131,6 +131,21 @@ def not_found?
       404 == @code
     end
 
+    #
+    # Base URI from the HTML doc head element
+    # http://www.w3.org/TR/html4/struct/links.html#edef-BASE
+    #
+    def base
+      @base = if doc
+        href = doc.search('//head/base/@href')
+        URI(href.to_s) unless href.nil? rescue nil
+      end unless @base
+
+      return nil if @base && @base.to_s().empty?
+      @base
+    end
+
+
     #
     # Converts relative URL *link* into an absolute URL based on the
     # location of the page
@@ -142,7 +157,7 @@ def to_absolute(link)
       link = URI.encode(URI.decode(link.to_s.gsub(/#[a-zA-Z0-9_-]*$/,'')))
 
       relative = URI(link)
-      absolute = @url.merge(relative)
+      absolute = base ? base.merge(relative) : @url.merge(relative)
 
       absolute.path = '/' if absolute.path.empty?
 

diff --git a/spec/fakeweb_helper.rb b/spec/fakeweb_helper.rb
@@ -15,6 +15,7 @@ def initialize(name = '', options = {})
       @hrefs = [options[:hrefs]].flatten if options.has_key?(:hrefs)
       @redirect = options[:redirect] if options.has_key?(:redirect)
       @auth = options[:auth] if options.has_key?(:auth)
+      @base = options[:base] if options.has_key?(:base)      
       @content_type = options[:content_type] || "text/html"
       @body = options[:body]
 
@@ -33,7 +34,11 @@ def auth_url
     private
 
     def create_body
-      @body = "<html><body>"
+      if @base
+        @body = "<html><head><base href=\"#{@base}\"></head><body>"
+      else
+        @body = "<html><body>"
+      end
       @links.each{|l| @body += "<a href=\"#{SPEC_DOMAIN}#{l}\"></a>"} if @links
       @hrefs.each{|h| @body += "<a href=\"#{h}\"></a>"} if @hrefs
       @body += "</body></html>"

diff --git a/spec/page_spec.rb b/spec/page_spec.rb
@@ -126,5 +126,52 @@ module Anemone
       end
     end
 
+    it "should detect, store and expose the base url for the page head" do
+      base = "#{SPEC_DOMAIN}path/to/base_url/"
+      page = @http.fetch_page(FakePage.new('body_test', {:base => base}).url)
+      page.base.should == URI(base)
+      @page.base.should be_nil
+    end
+
+    it "should have a method to convert a relative url to an absolute one" do
+      @page.should respond_to(:to_absolute)
+
+      # Identity
+      @page.to_absolute(@page.url).should == @page.url
+      @page.to_absolute("").should == @page.url
+
+      # Root-ness
+      @page.to_absolute("/").should == URI("#{SPEC_DOMAIN}")
+
+      # Relativeness
+      relative_path = "a/relative/path"
+      @page.to_absolute(relative_path).should == URI("#{SPEC_DOMAIN}#{relative_path}")
+
+      deep_page = @http.fetch_page(FakePage.new('home/deep', :links => '1').url)
+      upward_relative_path = "../a/relative/path"
+      deep_page.to_absolute(upward_relative_path).should == URI("#{SPEC_DOMAIN}#{relative_path}")
+
+      # The base URL case
+      base_path = "path/to/base_url/"
+      base = "#{SPEC_DOMAIN}#{base_path}"
+      page = @http.fetch_page(FakePage.new('home', {:base => base}).url)
+
+      # Identity
+      page.to_absolute(page.url).should == page.url
+      # It should revert to the base url
+      page.to_absolute("").should_not == page.url
+
+      # Root-ness
+      page.to_absolute("/").should == URI("#{SPEC_DOMAIN}")
+
+      # Relativeness
+      relative_path = "a/relative/path"
+      page.to_absolute(relative_path).should == URI("#{base}#{relative_path}")
+
+      upward_relative_path = "../a/relative/path"
+      upward_base = "#{SPEC_DOMAIN}path/to/"
+      page.to_absolute(upward_relative_path).should == URI("#{upward_base}#{relative_path}")      
+    end
+
   end
 end