I guess the bottom line is that we either need to accept "regressions"
there or come up with a more reliable way to spot these "regressions"
On Wed 25 Jul 2012 09:19:12 AM CDT, Steve Ebersole wrote:
Moving population of the JPA metamodel is the most likely culprit as
I
think you already knew.
The problem as we also already discussed, imo, is the test set up on
metamodel branch. For even minor changes you really have to run the
full test suite 3 times and keep track of the number of tests run and
passed/failed between 2 of the runs and apply a diff on the test
reports between those runs as well. This is just not reasonable. I
should be able to just run a testsuite and immediately know if
something I changed broke something else. We discussed scripting the
many-step manual process you do now into the gradle scripts.
On 07/25/2012 01:07 AM, Gail Badner wrote:
> I'm seeing 2 test failures using the new metamodel for tests that
> were passing in commit 787ab27d459a385411d3bf6ae0adfb9682898aa1.
>
> The tests are:
>
> org.hibernate.test.entityname.EntityNameFromSubClassTest
> org.hibernate.test.iterate.IterateTest
>
> Both of these tests are mapped using hbm.xml with the entity name
> specified. For EntityNameFromSubClassTest, the mapping is:
>
> <class name="Vehicle" abstract="true"
entity-name="VEHICLE" ...>
>
> The exception is:
>
> Caused by: org.hibernate.MappingException: Unknown entity:
> org.hibernate.test.entityname.Vehicle
>
> Anyone know of a change in the last couple of days that would cause
> this?
>
> Thanks,
> Gail
>
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/hibernate-dev
>