@@ -49,7 +49,7 @@ module Scrubyt debugger if options[:debug] if locator.is_a?(Array) if options[:compound] - all_matched_elements = locator.map{|l| parsed_doc.search(clean_xpath(l)).to_a} + all_matched_elements = locator.map{|l| evaluate_xpath(l).to_a} matching_elements = [] while(!all_matched_elements.empty?) do merged_element = Nokogiri::HTML.parse("") @@ -60,10 +60,10 @@ module Scrubyt all_matched_elements.reject!{|e| e.size < 1} end else - matching_elements = locator.map{|l| parsed_doc.search(clean_xpath(l))}.flatten + matching_elements = locator.map{|l| evaluate_xpath(l)}.flatten end else - matching_elements = parsed_doc.search(clean_xpath(locator)) + matching_elements = evaluate_xpath(locator) end return merge_elements(matching_elements, options[:script]) if merge_elements?(locator) matching_elements.each do |element| @@ -82,7 +82,43 @@ module Scrubyt ## Return only non-nil/empty results. Look at porting rails #blank? return @current_result = results.compact.reject{|r| r == ""} if options[:required] @current_result = results - end + end + + def evaluate_xpath(xpath) + if(xpath =~ /frameset/i) + xpath = clean_xpath(xpath) + ## split up the cleaned XPath to 2 parts: + ## - frame_path - just to grab the src attribute of the frame to navigate there + ## - element_path - the 'real' XPath pointing into the frame document + frame_path = xpath.scan(/.+frame.+?\]/)[0] + element_path = xpath[frame_path.size..-1] + + ## go there + ## TODO - how do we know whether we are still in that frame or not?!!?!?!?!?! + ## if we navigate away to a detail page, that's cool - however, what about the case + ## + ## parent1 ----> frames + ## - child1 + ## - child2 + ## parent2 ----> this is *not* in a frame, so the original doc should be re-loade + ## - child 3 + ## + ## I think this case is highly unlikely to ever come up + ## and btw an easy workaround is to put parent2 before parent1 + ## but there might be another cases? + ## dunno and don't want to spend too much time w/ it now + p parsed_doc + p frame_path + frame_src = parsed_doc.search(frame_path)[0].attributes["src"] + fetch(frame_src) + reset_page_state! + + ## and evaluate the XPath + parsed_doc.search(element_path) + else + parsed_doc.search(clean_xpath(xpath)) + end + end def extract_detail(result_name, *args, &block) locators = args.shift @@ -91,7 +127,7 @@ module Scrubyt locators = [locators] end if args.include?({:compound => true}) - all_matched_elements = locators.map{|l| parsed_doc.search(clean_xpath(l)).to_a} + all_matched_elements = locators.map{|l| evaluate_xpath(l).to_a} matching_elements = [] while(!all_matched_elements.empty?) do merged_element = Nokogiri::HTML.parse("") @@ -114,7 +150,7 @@ module Scrubyt end else results = locators.map do |locator| - parsed_doc.search(clean_xpath(locator)).map do |element| + evaluate_xpath(locator).map do |element| options = @options options.delete(:json) options[:json] = args.detect{|h| h.has_key?(:json)}[:json].to_json if args.detect{|h| h.has_key?(:json)} lib/scrubyt/results_extraction.rb 811f4ec8ac6741bf164297fd908b3e4f0e3d4b67 Peter Szinek peter@rubyrailways.com http://github.com/scrubber/scrubyt/commit/215cf771b98e91bc27e3f65311effdfbee58a9b7 215cf771b98e91bc27e3f65311effdfbee58a9b7 2009-10-19T10:54:21-07:00 2009-10-19T10:54:21-07:00 Considering frames during XPath evaluation 67fb4ed8772dc654f6aeff7a8d620b02dcba8729 Peter Szinek peter@rubyrailways.com