Apache JDO Project - Minutes: JDO TCK Conference Call Friday, Mar 24, 9 am PST

Author

Minutes: JDO TCK Conference Call Friday, Mar 24, 9 am PST

Craig L Russell

2006-03-25, 9:46 am

Michelle Caisse

2006-03-25, 9:46 am

Matthew Adams also attended the meeting.

-- Michelle

Craig L Russell wrote:

> Attendees: Michael Bouschen, Michelle Caisse, Martin Zaun, Craig Russell
>
> Agenda:
>
> 1. Test status (Michael) 1239 test cases; 2 errors, 4 failures; 4 of
> 57 configurations fail. Still an open issue on JDO-273 to update the
> life cycle for serialization with an active transaction. AI Martin
> fix the lifecycle tests and update the JIRA issues with the latest
> results. JPOX has issues with lifecycle and newInstance
> (PersistentInterface).
>
> 2. New tests status (Martin, Michael, Craig) Martin has one more test
> case to write: PMF close. Craig is still working on the datastore
> mapping for identity vs. sequence for primary key generation.
>
> 3. JSR 243 next steps (Craig) Assertions spreadsheet needs to be
> updated to reflect the new tests. New and updated TCK tests need to
> be finalized, and then we can cut the release branch. You can see the
> latest changes at http://svn.apache.org/viewcvs.cgi/db/jdo/trunk.
> JPOX is preparing a 1.1.0-rc2 for early next week, with the final
> 1.1.0 release scheduled for next Friday. The final artifact-id is
> expected to be jpox-1.1.0.
>
> 4. Other issues (any and all)
>
> Craig has instructions on creating a release, and will add a wiki
> page with the instructions in case anyone else has an itch to make a
> release.
>
> Need to resolve serialization of a detachable instance for durable
> storage. AI Craig: propose a solution for the specification.
>
> Action Items from weeks past:
>
> [Mar 17 2006] AI Craig: review EG discussion on JDOHelper
> constructor vs. newInstance(). Done. "There's no state in a JDOHelper
> instance so there's really not much of a reason to use a singleton
> pattern."
>
> [Mar 17 2006] Remaining issue: datastore identity strategy is
> currently defined as IDENTITY but some datastores don't support
> IDENTITY columns. AI CraIg: propose a solution. In progress
>
> [Mar 17 2006] For the persistence manager newInstance(Class) method,
> what should the implementation do for abstract classes or interfaces
> that contain abstract methods other than property accessors? Answer:
> throw an exception. This might happen during enhancement. This needs
> to be clarified in the spec. AI Michael: file a JIRA. AI Craig:
> update the change log. Done.
>
> [Feb 24 2006] One of the query tests fails. JDO-310 DERBY-39 for
> details. AI Craig: Can we get some help on this from the Derby
> team? Craig asked at the team meeting. AI Craig: follow up.
>
> [Feb 3 2006] AI Michael W fix list of teardown classes in company
> model reader. JDO-285. But this fix cannot be checked in until JPOX
> is fixed.
>
> [Jan 20 2005] AI Craig look at new life cycle tests to see if it
> can fail at the end printing failure messages. AI Craig review the
> spec updates to make sure the test is consistent. AI Craig:
> resolve the test case and spec issues. Done. Martin will start
> with the code that Michael W updated and will update the test case
> not to fail on the first error.
>
> [Nov 18 2005] AI: BEA to sign the donation paperwork for their test
> suite. Review how to merge their test cases into JDO TCK. The
> latest news in BEA is that BEA plans to donate their object
> persistence engine to open source; it's not clear where they will
> do this, perhaps in Apache as a sister project to Apache JDO.
> BEA's OpenJPA is currently under consideration in the Apache
> incubator, so the integration work might be facilitated once this
> is resolved.
>
> [Nov 4 2005] AI Martin: Update Martin's wiki with discusion of JDK
> 1.5 issues. in progress
>
> [Sep 2 2005] AI: To recruit members, update the web site. Articles
> on TheServerSide directing attention to the site. T-shirts, logo.
> AI: Craig write a ServerSide article.
>
> [July 29 2005] AI: Michelle Chapter 18 wiki needs to be updated to
> include all JDO metadata elements and attributes.
>
>
> Craig Russell
> Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
> 408 276-5638 mailto:Craig.Russell@sun.com
> P.S. A good JDO? O, Gasp!
>

Andy Jefferson

2006-03-27, 7:47 am

> 3. JSR 243 next steps (Craig). New and updated TCK tests need to
> be finalized, and then we can cut the release branch.
> JPOX is preparing a 1.1.0-rc2 for early next week, with the final
> 1.1.0 release scheduled for next Friday. The final artifact-id is
> expected to be jpox-1.1.0.

Is there going to be a jdo2-api 2.0-rc2 release ? If so, when ?
There was a method signature change in FetchPlan (getGroups) since rc1, and JPOX
1.1.0-rc-2 will be dependent on it.

--
Andy

Craig L Russell

2006-03-27, 7:47 am

Andy Jefferson

2006-03-27, 7:47 am

Hi Craig,

> Since tck20 final depends on JPOX which depends on api20, maybe it
> makes more sense to stage all three together. Is the intermediate
> JPOX rc2 really necessary?

In my opinion it is necessary. JPOX 1.1.0-rc-2 has 30 JIRA issues fixed relative
to 1.1.0-rc-1. This is a significant number and having a release very soon would
mean that users have a few days to find issues or otherwise 1.1.0-final will be
released. I would stress that I am keen to get 1.1.0-final out ASAP. The only
remaining issue is persistent interfaces and hopefully with Erik's recent changes
we should have that done in a day or so.

Of course JPOX could do what we have done in the past and just release JPOX
1.1.0-rc-2 with a jdo2-snapshot.jar (downloadable from our SourceForge site) that
it should be used with (we did this with the 1.1.0-rc-1 release so that people
had something to work with for the few days before jdo2-api-2.0-rc1.jar
appeared). Since the timeframe between JPOX 1.1.0-rc-2 and 1.1.0-final is likely
only a week then doing this may be the most effective way. What do you think ?

--
Andy