Bill Petheram <[EMAIL PROTECTED]> writes: > On Wednesday 22 January 2003 2:19 pm, [EMAIL PROTECTED] wrote: > > I agree the off should NOT be the default. Maven should encourage good > > practices, just not mandate them. Also, I think that projects like > > commons-collections have many test that run quickly because they don't have > > slow aspects like lots of filesystem access or testing parsing 1000 MB > > files, or interfacing with databases. But many business specific code does > > that type of thing, versus more general libraries. > > > > IMHO what you mention above are not unit tests. A Unit test should only test > one class. I know it is difficult to test classes that use databases and > sockets etc but either mockobjects should be used or move these sort of tests > out into an intgration test.
I agree, they aren't unit tests ( or level 1? ). I usually start all my tests at this level. If they become too slow, or need more external things than are available or something they get moved out to a different ring/level. Keeping the test fast and as close as possible to the code is what we strive for though. > Here we use a testharness package that is in src/java. This seems to work OK > but is not ideal. Like cactus for example. Your projects cactus integation tests execute after the normal build/unit test run. > > It sounds like consensus is that different levels would be the best, > > especially if you can still perform an on/off switch setup! > > > > Eric Pugh > > > > > > cheers > > bill > > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> -- ===================================================================== Jeffrey D. Brekke [EMAIL PROTECTED] Wisconsin, USA [EMAIL PROTECTED] [EMAIL PROTECTED] -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
