Overview
For Transaction tests, our agent waits for the Document Complete event before continuing each action listed under a step. This ensures that all referenced elements are present on a page. This behavior can cause problems on pages that use AJAX, as the agent is not loading a new page for each step. In these cases, the agent may end up waiting for a Document Complete that will never occur.
Workaround
To address this, we added a custom action "ignoreDocComplete()" (not present in Selenium)
Here is an example of an AJAX use case with Google search. When a user clicks the "Google Search" button the agent will load a new webpage, hence a "Document Complete" will fire.
However, when performing filtering on the results page -- say, viewing items from the "Past week" -- the page refreshes using AJAX, and there is no Document Complete event.
For this example, the search refinement results are AJAX, and the Transaction script looks like this:
` // Step - 1
open(http://www.google.com/)
type(//*[@id='lst-ib'], 'shoe')
// Step - 2
submit(//*[@id="gbqf"])
click(//*[@id='hdtb-tls'])
click(//*[@id="hdtbMenus"]/div/div[2]/div)
ignoreDocComplete()
// Step - 3
clickAndWait(link=regexp:Past hour)
ignoreDocComplete(
waitForUrl("www.google.com/search") `
The ignoreDocComplete action is used in conjunction with a "waitFor" action, such as waitForElementPresent or waitForUrl or waitForNoRequest, so the agent has something other than document complete to indicate that it can continue with the next action of the step.