On Wed, 19 Jan 2011 17:23:14 +0100, Emmanuel Bernard <emmanuel(a)hibernate.org&gt; wrote: > AFAIK we stayed on JUnit 3.8 to keep 1.4 compatibility. > Now that this is gone, what do you think about moving to JUnit 4.8 I think moving to JUnit 4.8 shouldn't be hard, especially since it contains a Junit 3 runner so that you don't have to convert all test in one go. If I remember correctly there was always another reason except the Java 1.4 compatibility. I think it had to do with how test classes are instantiated. For this reason we also discussed once testNG (which also had some shortcomings) > There are a few interesting new features besides the annotation goodness. > I am particularly interested in Rules which is essentially behavioral injection before / after test > http://kentbeck.github.com/junit/doc/ReleaseNotes4.7.html I've seen a presentation once about someone using rules and it looked interesting. As said, I don't think it should be hard to switch to JUnit 4. However, the more specific features we are going to use the harder a potential switch of frameworks become. Maybe Steve can fill us in again which particular problem there was with JUnit. --Hardy

Forgot to mention, Envers is using testNG. It would be nice to align frameworks. On Wed, 19 Jan 2011 17:23:14 +0100, Emmanuel Bernard <emmanuel(a)hibernate.org&gt; wrote: > AFAIK we stayed on JUnit 3.8 to keep 1.4 compatibility. > Now that this is gone, what do you think about moving to JUnit 4.8 > > There are a few interesting new features besides the annotation goodness. > I am particularly interested in Rules which is essentially behavioral > injection before / after test > http://kentbeck.github.com/junit/doc/ReleaseNotes4.7.html > > What do you think? > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

> I have tried testng in the past and was not overly happy there either. > If I remember correctly the issue there was an inability to alter test > outcome (altering failures to success due to @FailureExpected e.g.). That seems to be supported in TestNG: @Test(expectedExceptions = "...")