Maybe you hit that one: "Failsafe project does not fail in verify
phase when a test case object errors during initialization" [1]. Fixed
in 2.19.
--Gunnar
[1]
Concerning the original error, we are bitten by the changed version
number scheme in JDK 9 (we evaluate the Java version to enable
specific code when running on JDK 8 or later): We are expecting a
dot-separated version string such as "1.8.0", but JDK 9 returns only
"9" as proposed by JEP 223 [1].
I've filed HV-1048 [2] for that.
Regarding the Maven/Jenkins issue I am not fully sure, the FailSafe
plug-in detects "There are test failures" during the verify phase but
for some reason it doesn't fail the build. It's an issue on the Maven
side, not Jenkins; Test failure just mark a build (of Maven job type)
as unstable (yellow), not as failed (red). Maybe you try the latest
FailSafe version (2.19.1)?
--Gunnar
[1]
http://openjdk.java.net/jeps/223
[2]
https://hibernate.atlassian.net/browse/HV-1048
2016-01-20 11:36 GMT+01:00 Sanne Grinovero <sanne(a)hibernate.org>:
> I'm puzzled.
> There seems to be a javax.validation related issue which is failing
> the integration tests running on Arquillian / WildFly for the
> Hibernate Search / WildFly tests on JDK9.
>
>
http://ci.hibernate.org/view/JDK9/job/hibernate-search-master-jdk9/291/co...
>
> But more importantly.. anyone has a clue on why the build is marked as
> "SUCCESS" ?
>
> You actually have to notice the failures in the logs to figure out
> that it's not successful at all;
> search for "java.lang.ArrayIndexOutOfBoundsException" in the console logs.
>
> Sanne
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/hibernate-dev