We *really* need a development model change !

[Andriy Palamarchuk]

Alexandre Julliard wrote:

> I don't really care about the output. My point is
that this output is
> not necessary, a simple "assertion foo failed at
line xx" on failure
> and no output at all on success would work just as
well. I've nothing
> against such an output either, but I don't think it
justifies
> introducing the complexity of reusing the Test
stuff.

I already created small library with api like in
Test::Simple - functions ok, TODO, SKIP. The
implementation can be easy changed to the behaviour
you described. I also suggest for library to fail on
successful TODO test. TODO tests will be conditional -
they are TODO only under Wine, not on Windows.

Careful usage SKIP sections can easy cover differences
between Windows platforms, using TODO - between
different non-Microsoft Win32 implementations.

On other hand, SKIP sections were important for
Test::Harness, which needed to control number of
tests.
For your goals conditional execution of the tests is
sufficient.

Which approaches can you suggest to reach goals I
believe are important:

1) bundling - are we going to have separate
distributions - Wine with tests, tests only, Wine
only? There are a lot of cases where only one of them
is required.

2) development of the unit tests under Windows.
Obviously, we don't need to have Wine itself when we
work with unit tests on Windows. Plus, we need to
create development environment, usable by Windows
developers.

3) Organization of the unit tests in such way, so they
can be used by other Win32 implementation projects.
Conditional TODOs I suggested above will help to
manage different TODO lists for different projects.

4) besides narrow-scope regression testing you
suggested we need to run all the tests from time to
time, even if it takes more than 1 hour to run the
whole suite. The reasons for this are:

a) we need to run the tests on all Windows platforms
to make sure the tests correctly document their
behaviour

b) Wine should be reguraly tested by the whole testing
suite. Some regressions will be caught in unit tests
of the modules other than changed. In other cases
changed module may not have tests at all, but will
brake modules, dependent on it.

Thanks,
Andriy Palamarchuk

__________________________________________________
Do You Yahoo!?
Send FREE video emails in Yahoo! Mail!
http://promo.yahoo.com/videomail/