Vinktank

For those that have worked on an Agile project most would agree that integrating automated test cases with continuous integration is imperative. It provides regular feedback not only on the application but also on the tests themselves. Small iterative changes are easier to manage then large updates. There are challenges keeping the build time reasonable as well as ensuring the environment is up to date and kept in line with the latest code, but that is for another entry. Here I will talk about how to integrate Quick Test Pro (QTP) with Cruise Control. As tests written in Quick Test Pro quickly get to the stage where they are taking hour(s) to run I suggest you create a separate test build to the development (code) build.

There are different ways to integrate QTP and Cruise but this is one I first used with Winrunner. It does require another Mercury product, being Quality Center (or Test Director) and a little publicized tool called TDTestSetRun, which you can get from Mercury’s support site. This little executable allows you to call a test set from within Quality Center. Therefore you add all of the QTP tests you want to run into a test set in Quality Center (or you could do multiple and make multiple calls), then use TDTestSetRun to remotely invoke the test set. The trick is how to report back to Cruise that the tests failed (they will pass by default). To do this I used ant to check for the existence of a file and on each test in the Test Set created an on error (via preferences) task to create the file, which it will do only if the test fails.

There are of course drawbacks to this approach (apart from the cost of the tools!), for example the test results are not communicated back to Cruise (only a pass/fail result). By using Quality Center’s API you could get more information out if you wanted to. It will only run from a Windows PC and the TDTestSetRun application is very buggy. It gives very little and poor feedback if you make a mistake and seems to be very sensitive about the name of the folders and test set’s. One tip if you are having problems is make the test set appear at the top, so call it automated tests for example. Also I have found using a space in the test set name and not in the folder name works.

Below is a sample build.xml file which demonstrates a call to TDTestSetRun and fails if a file exists.

<project name=”Example” default=”runtest” basedir=”.”>
<description>
simple example build file for calling QTP or Winrunner tests in Quality Center
</description>
<!– set global properties for this build –>
<property name=”TDRunTestDir” location=”C:\TDTestSetRun”/>

<target name=”init”>
<delete file=”${TDRunTestDir}\TestFailed.txt”/>
</target>

<target name=”runtest” depends=”init”
description=”Run test set from Quality Center” >
<exec executable=”${TDRunTestDir}\TDRunTestSet.exe” output=”${TDRunTestDir}\Result.txt”>
<arg line=”/s:http://qulaitycenterserver/qcbin”/>
<arg line=”/n:Domain”/>
<arg line=”/d:Project Name”/>
<arg line=”/u:username”/>
<arg line=”/p:password”/>
<arg line=””/t:Automated Tests””/>
<arg line=”/f:Root\Folder\”/>
<arg line=”/l”/> <!–This option is to run the QTP tests locally.–>
</exec>
<fail message=”Tests Failed, see Quality Center for more information”>
<condition>
<available file=”${TDRunTestDir}\TestFailed.txt”/>
</condition>
</fail>
</target>

</project>