On 2/24/12 9:33 AM, Raphael Bircher wrote:
Am 24.02.12 09:06, schrieb Jürgen Schmidt:
On 2/24/12 8:33 AM, O.Felka wrote:
Am 24.02.2012 06:20, schrieb Zhe Liu:
在 2012年2月24日 上午10:37，Ma Yong Lin<[email protected]> 写道：
I think what was asked is if we are just get a group of star basic
test cases "translated"
to java. It would be the same in terms of testing.
We can do that, but I think it's more meaningful to cover the GUI
testing for new functions in future, since the existing function has
been very stable.
I think that it is worth the effort and very helpfull to translate the
old Star Basic tests to Java. We have had always regressions in the
existing and 'stable' functions. So we really need the old tests for
finding regressions in the existing functionality.
every test if old or new working with this new automation tool is
welcome and highly appreciated. Maybe it's easier for new develoeprs to
work with existing tests and make them available for the new
environment. Volunteers are always welcome.
But in generaI would agree that we should focus on new tests for new
changes. And most people would like to work on something new ;-)
No, new feature are allways good tested. Everyone focus on new futers in
manual tests, and that is good so. Automated tests should be used to
garantee basic stability. You can run it 100 times without wasting too
much human ressource. So finely we should focuse on basic functionality
at auto test.
I understand that Raphael, the point is the interest of rewriting the
existing tests to the new tool goes probably to zero. I don't say that
we shouldn't do that but I won't expect to many volunteers here.
If somebody is interested to rewrite these tests I am sure she/he will
get all the support we can provide. And of course as I mentioned earlier
it would be highly appreciated.
And feel free to start working on it based on the new tooling.
Even you don't realy like the old TT. The Scripts are a good starting
point, because they cover a load of the basic functionality.
I never said something else. The point is simply that the old tool was
not maintained for a long time, we have a lack of knowledge and it is
probably wasted time to investigate in it further.
If the tool still work it's fine and if somebody would like to work on
it it's also fine. But I favor the new approach for reasons I have
already pointed out.
And when we can integrate such a tooling in the build process and do it
automatically it's even better ;-)
Juergen
Greetings Raphael
