Dain Sundstrom wrote:
This is great news :)
We can finally remove the lame hack in the CmpJpaEngine (or what ever
I named that class) that works around this bug. It is the code that
tracks the newly created instances so we don't have to ask the JPA code.
Yes, I've already removed that code and all of the tests appear to run
cleanly. I'll check that change in when I make the switch.
Rick
-dain
On Jul 10, 2008, at 11:10 AM, David Blevins wrote:
On Jul 10, 2008, at 6:36 AM, Rick McGuire wrote:
I was looking at the JpaTest test cases today that Dain recommended
yesterday, and noticed that there were a couple of asserts that were
commented out and marked with "FAILS" in the newDeleteNew() method.
As an experiment, I updated the build to OpenJPA 1.1.0, and enabled
those tests. They now pass on the most recent version of OpenJPA.
I suspect we might want to consider upleveling OpenJPA now. This
would allow some code in the CMP container that implemented a
work-around for this bug to be removed. Some of the other
workarounds since as marking the fields that OpenJPA adds as
transient might also be removable.
Sounds great to me.
-David
