hi john,

if the feature is enabled, the test-class gets a std. cdi-bean (lookup via
BeanManager).

regards,
gerhard

http://www.irian.at

Your JavaEE powerhouse -
JavaEE Consulting, Development and
Courses in English and German

Professional Support for Apache
MyFaces, DeltaSpike and OpenWebBeans

2017-11-26 18:53 GMT+01:00 John D. Ament <[email protected]>:

> Gerhard,
>
> RE the original question.  If a scope is applied to the test class, does it
> handle it correctly?  E.g. if its request scoped and each test method
> starts a new request context, is a new instance used?  Or is it treated as
> a dependent bean regardless?
>
> John
>
> On Mon, Nov 20, 2017 at 6:52 AM Gary Hodgson <[email protected]>
> wrote:
>
> > Hi Gerhard,
> >
> > I can understand the desire not to confuse the reader, but I think there
> is
> > more risk of confusion by not documenting this.  For example, I came
> across
> > the problem because we use Database Rider to setup DBUnit, and
> > documentation in this downstream project explicitly advises to set the
> > property to true (
> > https://database-rider.github.io/getting-started/#configuration_3)
> >
> > We now have the decision to either revert the use of Database Rider, or
> to
> > use it and accept, and document, the consequence that Junit setup
> > functionality is not available in those tests.  Having a warning in the
> > Test Control Module documentation would have allowed us to make this
> > decision earlier.
> >
> > In any case I will suggest to the Database Rider author that he put in a
> > warning in his documentation as I think it is an important point to
> > mention.
> >
> >
> > Regards,
> > Gary
> >
> > On 20 November 2017 at 12:23, Gerhard Petracek <[email protected]>
> > wrote:
> >
> > > hi gary,
> > >
> > > that flag shouldn't be in the documentation at all (to avoid
> confusion).
> > > we just kept the possibility to use test-classes as cdi-beans, because
> it
> > > was the default behavior prior v1 (before we found the issue with
> > @Before)
> > > and some users used it extensively.
> > >
> > > regards,
> > > gerhard
> > >
> > >
> > >
> > > 2017-11-20 11:14 GMT+01:00 Gary Hodgson <[email protected]>:
> > >
> > > > Hi Gerhard,
> > > >
> > > > Thanks for confirming that.
> > > >
> > > > I couldn't find mention of this in the documentation (perhaps it
> passed
> > > me
> > > > by), perhaps it would be worthwhile adding a sentence to the Test
> > Control
> > > > Module for others in the future?
> > > >
> > > > All the best,
> > > > Gary
> > > >
> > > > On 19 November 2017 at 23:24, Gerhard Petracek <
> > > [email protected]
> > > > >
> > > > wrote:
> > > >
> > > > > hi gary,
> > > > >
> > > > > exactly that issue was the reason for the flag and the reason why
> > it's
> > > > > false per default.
> > > > >
> > > > > regards,
> > > > > gerhard
> > > > >
> > > > > http://www.irian.at
> > > > >
> > > > > Your JavaEE powerhouse -
> > > > > JavaEE Consulting, Development and
> > > > > Courses in English and German
> > > > >
> > > > > Professional Support for Apache
> > > > > MyFaces, DeltaSpike and OpenWebBeans
> > > > >
> > > > >
> > > > >
> > > > > 2017-11-19 21:07 GMT+01:00 Gary Hodgson <[email protected]
> >:
> > > > >
> > > > > > Hi
> > > > > >
> > > > > > I have a question about using @Before in CdiTestRunner enabled
> unit
> > > > > tests.
> > > > > >
> > > > > > The following test fails when I set the config property
> > > > > > deltaspike.testcontrol.use_test_class_as_cdi_bean=true
> > > > > >
> > > > > >   @RunWith(CdiTestRunner.class)
> > > > > >   public class ATest {
> > > > > >
> > > > > >       boolean a;
> > > > > >
> > > > > >       @Before
> > > > > >       public void setup() {
> > > > > >           a = true;
> > > > > >       }
> > > > > >
> > > > > >       @Test
> > > > > >       public void testA() {
> > > > > >           assertTrue(a);
> > > > > >       }
> > > > > >   }
> > > > > >
> > > > > >
> > > > > > Is this expected behaviour because the test class is processed
> as a
> > > CDI
> > > > > > bean? Or is it a bug?
> > > > > >
> > > > > > Cheers,
> > > > > > Gary
> > > > > >
> > > > >
> > > >
> > >
> >
>

Reply via email to