Desktop automated tests

Registered by Henrik Nilsen Omma

Leverage the ATK framework to implement a set of automated test routines for the desktop.

Agenda:

 * Choice of technology - Evaluate Accerciser, Macaroon and Dogtail. Criteria: ease of use, availability (is dogtail still broken?), compatibility with other test projects

 * Test structure - how long should test be?; how should multiple tests be chained together?

 * Generating tests:
  * Identify 5 target apps from the default desktop and record tests and clean up results as needed. Document the steps involved.
  * Encourage community contributions following the established procedure
  * Investigate auto-generated tests (model-based or random walk)

 * Test output
  * What should the result of a test be and how should it be evaluated for a pass/fail rating?
  * Can we test for proper functionality at this point, or should that be phase 2?

 * Policy for work-arounds
  * Some applications may require a work-around if ATK support is incomplete - such limitations shuld be noted, a bug should be filed and tracked

Blueprint information

Status:
Started
Approver:
Henrik Nilsen Omma
Priority:
High
Drafter:
None
Direction:
Approved
Assignee:
Ara Pulido
Definition:
Approved
Series goal:
Accepted for intrepid
Implementation:
Beta Available
Milestone target:
None
Started by
Henrik Nilsen Omma

Related branches

Sprints

Whiteboard

2007-10-09 kamion: This could also cover automated testing of desktop installs using ubiquity's automation framework and maybe vmware.
2007-11-16: Changing priority to High; it's important but it's not a critical part of the release itself
2008-06-12 heno: Assigning to Ara and targeting for Intrepid

(?)

Work Items

Dependency tree

* Blueprints in grey have been implemented.