You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jdo-dev@db.apache.org by Michelle Caisse <Mi...@Sun.COM> on 2005/08/18 23:41:53 UTC
JDO TCK Conference Call Friday, Aug 19, 9 am PST
Hi,
We will have our regular meeting Friday, August 12 at 9 am PST to
discuss JDO TCK issues and status.
Dial-in numbers are:
866 230-6968 294-0479#
International: +1 865 544-7856
Agenda:
Test status (Michael, Michelle)
JPOX fixes/issues (Erik)
Detached objects (Matthew)
Other issues and status (any and all)
Action Items from weeks past:
[Aug 12 2005] AI: Craig to propose release of API20 and the entire 11
release (API, RI, TCK).
Other issues and status (any and all)
<>[Aug 12 2005] AI: Michael B reply to the email and start the dialog.
re: Should we do cleanup in teardown or cleanup? Brian suggests doing
database cleanup during setUp; this has advantages that we should
explore. Perhaps a system property for CleanupAtSetup and another
CleanupAtTeardown?
[Aug 5 2005] AI Brian McCallister can send info on the instructions how
to sync Apache and ibiblio. Re: Brian Topping needs info on creating a
maven package.<>
[Aug 5 2005] AI: Karan: Issue with maven: renaming JPOX releases to
SNAPSHOT. Need to update the README file and project.xml file.
[Aug 5 2005] AI: Geoff look into alternative ways to get this project.
Re: Maven needs to checkout files from netbeans, not documented well.
[July 29 2005] AI: Michelle Chapter 18 wiki needs to be updated to
include all JDO metadata elements and attributes. [not done]
[July 29 2005] Inheritance proposal sent around by Michael. Needs to be
reviewed. 5 inheritance strategies. AI: Craig forward proposal to expert
group.
[July 29 2005] Is there any way to indicate from maven whether a set of
configurations was successful? Not the way it's set up. AI: Michael file
a JIRA on this.
[July 29 2005] (Craig, Brian T.) Need some permissions that Brian is
working on.
[July 22 2005] ? We need to enhance the DeepEquality class to allow
reporting of the difference reported. This will allow easier debugging
of mapping issues. Craig filed JIRA issue 109 on this.
[Jul 8] Double-check locking in the PMF (Martin) Martin has implemented
and will check in.
[April 22 2005] AI JDO API release on ibiblio (Brian T)
[May 20 2005] AI: Craig file JIRA bug and fix this. Can test cases rely
on JDOHelper for state interrogation of non-binary-compatibility
classes? Yes, but the JDO 2.0 JDOImplHelper class has not implemented
the functionality yet.
[April 15 2005] AI Brian Topping will update the wiki to tell how to
access our releases area.
[April 15 2005] AI Brian Topping will do the maven goal for creating and
uploading the snapshots. He will create a directory parallel to trunk
called "releases" and put the snapshots there.
[April 15 2005] AI Matthew will create a directory in the
tck.api.persistencemanager called detach in which he will have complete
freedom to implement the assertions in the detach section of the
specification.
[May 13 2005] AI Brian Topping will implement pushing SNAPSHOT builds of
the project to ibiblio.
[May 13 2005] AI Brian Topping will arrange for automated nightly builds.
[May 13 2005] AI Martin Zaun will investigate JSR 294 (Java 5) to see
impact on enhancer.
[May 20 2005] AI Craig to define the JCP distributions and see if maven
can help.
-- Michelle
Re: Minutes: JDO TCK Conference Call Friday, Aug 19, 9 am PST
Posted by Andy Jefferson <an...@jpox.org>.
Hi Craig,
> > Feedback : CollectionCollections has a field Object1 that is a
> > Collection<SimpleClass> embedded in the join table. The TCK is
> > expecting a
> > schema with join table having columns for the FK back to the owner
> > object,
> > and the fields for the element. JPOX looks at this from a different
> > angle and
> > requires a PK on this join table and so, since we don't have an id
> > for the
> > embedded SimpleClass, we add an "adapter index" column to form the
> > PK. As has
> > been discussed in the past, the user has no obvious way of saying
> > that they
> > don't want a PK applying to the join table. They have the <primary-
> > key>
> > element but that is only currently for specifying the pk name.
>
> There is a natural PK for the join table. It's a combination of the
> FK plus the column mapped to the id field. If this were defined to be
> the PK for the join table in the schema, would that work?
If the implementation can be handed a definition of the columns to use in the
PK for the join table, then the implementation can use that, and saves it
having to resort to adding its own magic - which would be a good thing from a
TCK perspective.
In this particular case as you say, the SimpleClass has a field "id" that is
persistent, and so will be present in the join table whilst embedded and, as
long as its value is consistent with use as part of the PK, could be used as
such.
Looking at the PFD 18.7, I see you've added that the user can specify
<join table="...">
<primary-key name="...">
<column name="..."/>
<column name="..."/>
</primary-key>
</join>
You'll just have to give me time to implement this ;-). JPOX supports the
<primary-key name="..."/> syntax but not the subelement columns yet.
Section 18.7 doesn't mention what happens if the user doesn't specify this PK
definition, consequently JPOX will just define the PK how it does at the
moment if nothing is provided.
--
Andy
Java Persistent Objects - JPOX
Re: Minutes: JDO TCK Conference Call Friday, Aug 19, 9 am PST
Posted by Craig Russell <Cr...@Sun.COM>.
Hi Andy,
On Aug 20, 2005, at 2:25 AM, Andy Jefferson wrote:
>> CollectionOfObject doesn't work JDO-121. AI: Need feedback from Erik
>> or Andy.
>>
>
> Feedback : CollectionCollections has a field Object1 that is a
> Collection<SimpleClass> embedded in the join table. The TCK is
> expecting a
> schema with join table having columns for the FK back to the owner
> object,
> and the fields for the element. JPOX looks at this from a different
> angle and
> requires a PK on this join table and so, since we don't have an id
> for the
> embedded SimpleClass, we add an "adapter index" column to form the
> PK. As has
> been discussed in the past, the user has no obvious way of saying
> that they
> don't want a PK applying to the join table. They have the <primary-
> key>
> element but that is only currently for specifying the pk name.
There is a natural PK for the join table. It's a combination of the
FK plus the column mapped to the id field. If this were defined to be
the PK for the join table in the schema, would that work?
Craig
>
> In addition, since the test is using java.util.Collection, which
> can allow
> duplicates, the JDO impl can utilise an "adapter index" in the same
> way to
> allow duplicates in the join table.
>
>
>
> --
> Andy
> Java Persistent Objects - JPOX
>
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!
Re: Minutes: JDO TCK Conference Call Friday, Aug 19, 9 am PST
Posted by Andy Jefferson <an...@jpox.org>.
> CollectionOfObject doesn't work JDO-121. AI: Need feedback from Erik
> or Andy.
Feedback : CollectionCollections has a field Object1 that is a
Collection<SimpleClass> embedded in the join table. The TCK is expecting a
schema with join table having columns for the FK back to the owner object,
and the fields for the element. JPOX looks at this from a different angle and
requires a PK on this join table and so, since we don't have an id for the
embedded SimpleClass, we add an "adapter index" column to form the PK. As has
been discussed in the past, the user has no obvious way of saying that they
don't want a PK applying to the join table. They have the <primary-key>
element but that is only currently for specifying the pk name.
In addition, since the test is using java.util.Collection, which can allow
duplicates, the JDO impl can utilise an "adapter index" in the same way to
allow duplicates in the join table.
--
Andy
Java Persistent Objects - JPOX
Re: Minutes: JDO TCK Conference Call Friday, Aug 19, 9 am PST
Posted by Michelle Caisse <Mi...@Sun.COM>.
You're right, Andy. I will make this change.
-- Michelle
Andy Jefferson wrote:
>>FieldsOfObject still doesn't work. JDO-120. AI: Need feedback from
>>Erik or Andy.
>>
>>
>
>Feedback : TestFieldsOfObject.java is not correct. It's still trying to use
>Object values that are String, and Integer even though the metadata is now
>specifying that the "implementation" of the Object field is SimpleClass. The
>"firstValue", "secondValue" values need changing to be SimpleClass objects.
>
>JPOX is throwing a ClassCastException as is required by the spec 6.4.3
><spec>
>If an implementation restricts instances to be assigned to the field, a
>ClassCastException must be thrown at the time of any incorrect assignment.
></spec>
>though the message currently given in the exception could be made more
>explicit.
>
>
>
>
Re: Minutes: JDO TCK Conference Call Friday, Aug 19, 9 am PST
Posted by Andy Jefferson <an...@jpox.org>.
> FieldsOfObject still doesn't work. JDO-120. AI: Need feedback from
> Erik or Andy.
Feedback : TestFieldsOfObject.java is not correct. It's still trying to use
Object values that are String, and Integer even though the metadata is now
specifying that the "implementation" of the Object field is SimpleClass. The
"firstValue", "secondValue" values need changing to be SimpleClass objects.
JPOX is throwing a ClassCastException as is required by the spec 6.4.3
<spec>
If an implementation restricts instances to be assigned to the field, a
ClassCastException must be thrown at the time of any incorrect assignment.
</spec>
though the message currently given in the exception could be made more
explicit.
--
Andy
Java Persistent Objects - JPOX
Minutes: JDO TCK Conference Call Friday, Aug 19, 9 am PST
Posted by Craig Russell <Cr...@Sun.COM>.
Attendees: Michelle Caisse, Michael Bouschen, Michael Watzek, Martin
Zaun, Karan Mahli, Craig Russell
Agenda:
1. Test status (Michael, Michelle) Application identity 396 tests 17
failures 24 errors Datastore identity 16 failures 25 errors. Open
JIRA bugs: 17.
FieldsOfObject still doesn't work. JDO-120. AI: Need feedback from
Erik or Andy.
CollectionOfObject doesn't work JDO-121. AI: Need feedback from Erik
or Andy.
Logging patch: Aggregating results into a summary file. Needs just a
bit more review. The idea is to collect for all configurations how
many tests, errors, and failures into one file so users don't have to
grep through a dozen files and do the addition manually.
2. JPOX fixes/issues (Erik) no update.
3. Detached objects (Matthew) no update.
Other issues and status (any and all)
Action Items from weeks past:
[Aug 12 2005] AI: Craig to propose release of API20 and the entire 11
release (API, RI, TCK). This generated a large response on the
incubator alias.
Other issues and status (any and all)
<>[Aug 12 2005] AI: Michael B reply to the email and start the
dialog. re: Should we do cleanup in teardown or cleanup? Brian
suggests doing database cleanup during setUp; this has advantages
that we should explore. Perhaps a system property for CleanupAtSetup
and another CleanupAtTeardown?
Michael B sent a proposal for this. Awaiting responses.
[Aug 5 2005] AI Brian McCallister can send info on the instructions
how to sync Apache and ibiblio. Re: Brian Topping needs info on
creating a maven package.<>
[Aug 5 2005] AI: Karan: Issue with maven: renaming JPOX releases to
SNAPSHOT. Need to update the README file and project.xml file.
[Aug 5 2005] AI: Geoff look into alternative ways to get this
project. Re: Maven needs to checkout files from netbeans, not
documented well.
[July 29 2005] AI: Michelle Chapter 18 wiki needs to be updated to
include all JDO metadata elements and attributes. [not done]
[July 29 2005] Inheritance proposal sent around by Michael. Needs to
be reviewed. 5 inheritance strategies. AI: Craig forward proposal to
expert group.
[July 29 2005] Is there any way to indicate from maven whether a set
of configurations was successful? Not the way it's set up. AI:
Michael file a JIRA on this.
[July 29 2005] (Craig, Brian T.) Need some permissions that Brian is
working on.
[July 22 2005] ? We need to enhance the DeepEquality class to allow
reporting of the difference reported. This will allow easier
debugging of mapping issues. Craig filed JIRA issue 109 on this.
[Jul 8] Double-check locking in the PMF (Martin) Martin has
implemented and will check in.
[April 22 2005] AI JDO API release on ibiblio (Brian T)
[May 20 2005] AI: Craig file JIRA bug and fix this. Can test cases rely
on JDOHelper for state interrogation of non-binary-compatibility
classes? Yes, but the JDO 2.0 JDOImplHelper class has not implemented
the functionality yet.
[April 15 2005] AI Brian Topping will update the wiki to tell how to
access our releases area.
[April 15 2005] AI Brian Topping will do the maven goal for creating and
uploading the snapshots. He will create a directory parallel to trunk
called "releases" and put the snapshots there.
[April 15 2005] AI Matthew will create a directory in the
tck.api.persistencemanager called detach in which he will have complete
freedom to implement the assertions in the detach section of the
specification.
[May 13 2005] AI Brian Topping will implement pushing SNAPSHOT builds of
the project to ibiblio.
[May 13 2005] AI Brian Topping will arrange for automated nightly
builds.
[May 13 2005] AI Martin Zaun will investigate JSR 294 (Java 5) to see
impact on enhancer.
[May 20 2005] AI Craig to define the JCP distributions and see if maven
can help.
-- Michelle
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!