Re: Junit (was Re: [Classpathx-discuss] Servlet API)

[David Brownell]

> >The only reason I mentioned junit is because it 
> >is the framework that I have the most experience 
> >with, but I wouldn't have any objection to something 
> >else...

Do you know of other realistic options?  A few years
back, I didn't know of any Java-specific test harnesses.
JUnit gets a lot of airtime lately.


> If you want to start writing test cases right away then please use
> JUnit but keep the test cases out of the Classpathx CVS.
>
> If you can wait till Dave has had a chance to tell us what he thinks
> about testing we can come to a group decision.

Testing is essential, under-resourced, under-valued, and something
I wish more developers paid attention to!  It'll be essential to the
long term success of ClasspathX; ad-hock usage-driven testing is
not a workable solution for any very large bodies of code.

Getting a good testing framework (organizational and technical)
is hard.  I'm not sure I'd necessarily see a problem with running
GPL'd tests over a non-GPL'd driver framework (JUnit); it'd be
much the same as running GCC over a proprietary OS (which
in this case happened to be open source).

On the other hand, I'm also not sure JUnit can necessarily be the
answer to all our testing framework needs.  I've not used it yet
myself, though I expect to try using it with some of the DOM tests
that are being developed.  After I've done that, then I'll likely
have more specific opinions ... :)

In cases where there's already an existing Free (or even Open)
test community we can leverage, I think we should do so.  If
one of those communities builds on JUnit, I'd hope that would
be workable.

- Dave