Make "should contain" matcher work with inner text instead of HTML

lib/webrat/core/matchers/have_content.rb

@@ -3,12 +3,31 @@ module Webrat
   
     class HasContent
       def initialize(content)
+        # Require nokogiri and fall back on rexml
+        begin
+          require "nokogiri"
+          require "webrat/nokogiri"
+        rescue LoadError => e
+          if require "rexml/document"
+            require "webrat/vendor/nokogiri/css"
+            warn("Standard REXML library is slow. Please consider installing nokogiri.\nUse \"sudo gem install nokogiri\"")
+          end
+        end
+        
         @content = content
       end
       
-      def matches?(element)
+      def matches?(stringlike)
-        element = element.body.to_s if element.respond_to?(:body)
+        if defined?(Nokogiri::XML)
-        @element = element
+          matches_nokogiri?(stringlike)
+        else
+          matches_rexml?(stringlike)
+        end
+      end
+    
+      def matches_rexml?(stringlike)
+        @document = rexml_document(stringlike)
+        @element = @document.inner_text
       
         case @content
         when String
@@ -18,16 +37,63 @@ module Webrat
         end
       end
     
+      def matches_nokogiri?(stringlike)
+        @document = nokogiri_document(stringlike)
+        @element = @document.inner_text
+      
+        case @content
+        when String
+          @element.include?(@content)
+        when Regexp
+          @element.match(@content)
+        end
+      end
+    
+      def rexml_document(stringlike)
+        stringlike = stringlike.body.to_s if stringlike.respond_to?(:body)
+        
+        case stringlike
+        when REXML::Document
+          stringlike.root
+        when REXML::Node
+          stringlike
+        when StringIO, String
+          begin
+            REXML::Document.new(stringlike.to_s).root
+          rescue REXML::ParseException => e
+            if e.message.include?("second root element")
+              REXML::Document.new("<fake-root-element>#{stringlike}</fake-root-element>").root
+            else
+              raise e
+            end
+          end
+        end
+      end
+      
+      def nokogiri_document(stringlike)
+        return stringlike.dom if stringlike.respond_to?(:dom)
+        stringlike = stringlike.body.to_s if stringlike.respond_to?(:body)
+        
+        case stringlike
+        when Nokogiri::HTML::Document, Nokogiri::XML::NodeSet
+          stringlike
+        when StringIO
+          Nokogiri::HTML(stringlike.string)
+        else
+          Nokogiri::HTML(stringlike.to_s)
+        end
+      end
+      
       # ==== Returns
       # String:: The failure message.
       def failure_message
-        "expected the following element's content to #{content_message}:\n#{@element.inner_text}"
+        "expected the following element's content to #{content_message}:\n#{@element}"
       end
 
       # ==== Returns
       # String:: The failure message to be displayed in negative matches.
       def negative_failure_message
-        "expected the following element's content to not #{content_message}:\n#{@element.inner_text}"
+        "expected the following element's content to not #{content_message}:\n#{@element}"
       end
     
       def content_message

spec/api/matchers_spec.rb

@@ -58,28 +58,20 @@ describe Webrat::Matchers do
     include Webrat::Matchers
     
     before(:each) do
-      @element = stub(:element)
+      @body = <<-EOF
-      @element.stub!(:inner_text).and_return <<-EOF
         <div id='main'>
           <div class='inner'>hello, world!</div>
         </div>
       EOF
-      
-      @element.stub!(:include?)
-      @element.stub!(:match)
     end
     
     describe "#matches?" do
       it "should call element#include? when the argument is a string" do
-        @element.should_receive(:include?)
+        @body.should contain("hello, world!")
-        
-        Webrat::Matchers::HasContent.new("hello, world!").matches?(@element)
       end
       
       it "should call element#match when the argument is a regular expression" do
-        @element.should_receive(:match)
+        @body.should contain(/hello, world/)
-        
-        Webrat::Matchers::HasContent.new(/hello, world/).matches?(@element)
       end
     end
       
@@ -100,9 +92,9 @@ describe Webrat::Matchers do
       
       it "should include the element's inner content" do
         hc = Webrat::Matchers::HasContent.new(/hello,\sworld!/)
-        hc.matches?(@element)
+        hc.matches?(@body)
         
-        hc.failure_message.should include(@element.inner_text)
+        hc.failure_message.should include("hello, world!")
       end
     end
   end