running ubuntu-desktop-testing directly from the branch instead of installing it system-wide

On https://wiki.ubuntu.com/Testing/Automation/Desktop/GettingStarted under "How to run the tests" it says: "Go the newly created folder and run ./bin/ubuntu-desktop-test. That will run ALL the available tests."
This is a good idea, especially when you are working on new features. However when running this in lp:ubuntu-desktop-testing I get:

markus@thekorn ~/devel/ubuntu-desktop-testing/trunk % ./bin/ubuntu-desktop-test -a gedit :(
Traceback (most recent call last):
  File "./bin/ubuntu-desktop-test", line 488, in <module>
    sys.exit(main())
  File "./bin/ubuntu-desktop-test", line 295, in main
    process_suite_file(suite_file, options.target, options.case)
  File "./bin/ubuntu-desktop-test", line 211, in process_suite_file
    run_suite_file(suite_file, log_file, cases)
  File "./bin/ubuntu-desktop-test", line 165, in run_suite_file
    runner = TestSuiteRunner(suite_file)
  File "./bin/ubuntu-desktop-test", line 414, in __init__
    mod = __import__(modname)
  File "./gedit/gedit_chains.py", line 8, in <module>
    from desktoptesting.gnome import GEdit
ImportError: No module named desktoptesting.gnome
1 markus@thekorn ~/devel/ubuntu-desktop-testing/trunk %

As a solution for this I suggest using a script like zc.buildout which creates an 'isolated', local environment for ubuntu-desktop-testing.