Sara Ford - Analyzing a Test Case Failure (Maddog Part 3)

Play Sara Ford - Analyzing a Test Case Failure (Maddog Part 3)
Sign in to queue


Now that the test case owner knows why the test case has failed, the information, usually in the form of a bug id, is entered into Maddog.  This tracking allows everyone, not just the test case owners, to know how many product bugs are affecting the current build.

In this clip, watch MHagman and SaraF analyze a test case failure in Maddog.

For more Devdiv updates check out our C9 site:
- josh



Download this episode

The Discussion

  • User profile image

    To create the automated tests, do you record and then playback mouse and keyboard events?  I was assuming you'd do it programmatically with COM or maybe send WM_COMMANDs.

    A while ago I had to automate the AOL dial-up window because it didn't use a standard RAS entry.  Instead I found the control IDs with Spy++ and used PostMessage and SendMessage to click the right buttons.  I guess you don't do it that way because your goal is to simulate an actual user?

  • User profile image

    Record and playback is definitely one way to automate tests.  Our primary method is to use a combination of MSAA and Win32 api calls, like sending messages to a control as you describe above.  We drive the UI (meaning we use the mouse and/or keyboard to click on menus, buttons, etc) as much as possible to simulate an actual user, and it does a really good job.

    Brian McMaster, one of our software testing architects (someone who designs our automation framework) wrote this whitepaper on Automating WinForms...

    Hope this helps!

Add Your 2 Cents