On Jan 26, 2008, at 5:47 PM, Jeffrey Harris wrote:
A lot of work in 2007 went into attempting to improve Chandler's
framework for creating automated tests. Unfortunately, QA's
assessment is that the script recording framework just isn't at a
level yet where it can offer a way to create automated tests more
easily than using the old functional test framework.
I think that the recorded scripts framework is working pretty well as
long as we "stay on the beaten path", i.e. record scripts with events
that have been tested before. Most of the problems arise from
recording scripts using events that have never been recorded before.
This resulted from our decision to focus on implemented the most
scripts in the least amount of time, postponing doing a careful
inventory of what events worked and what events didn't work. If I had
to hazard a guess of what fraction of events work, I'd say it's
greater than 90% -- although if just one event in 50 fails that can
lead to a test that fails. The list of which events work and which
don't still isn't well documented, but it could be without too much
more effort. So, from my perspective I think it's easier to create
automated tests than write tests using the old framework.
John
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev
