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 CR...@versant.com on 2009/07/30 13:27:52 UTC
broken tests in TCK testsuite
Hi all,
I hope this is the right list to report broken tests in the TCK testsuite,
I'm referring to TCK 2.0, ignore my post if these issues are resolved
meanwhile:
The following tests are broken, because they rely on a no-args ctor
requirement:
org.apache.jdo.tck.api.persistencemanagerfactory.SetConnectionPassword
org.apache.jdo.tck.api.persistencemanagerfactory.SetConnectionURL
org.apache.jdo.tck.api.persistencemanagerfactory.SetConnectionUserName
org.apache.jdo.tck.api.persistencemanagerfactory.SetIgnoreCache
org.apache.jdo.tck.api.persistencemanagerfactory.SetMultithreaded
org.apache.jdo.tck.api.persistencemanagerfactory.SetNonTransactionalRead
org.apache.jdo.tck.api.persistencemanagerfactory.SetNonTransactionalWrite
org.apache.jdo.tck.api.persistencemanagerfactory.SetOptimistic
org.apache.jdo.tck.api.persistencemanagerfactory.SetRetainValues
Per specification the instantiation of the vendor implementation of
javax.jdo.PersistenceManagerFactory is
implementation specific, so Class.newInstance() is not valid.
If there is any JDOHelper.getPersistenceManagerFactory() overload which
returns a mutable PMF, the tests
can be rewritten using this. Otherwise the tests can not be fixed.
Best regards,
Christian
---
Versant GmbH
European Headquarters
Christian Romberg
Wiesenkamp 22 b
D 22359 Hamburg
---
Versant GmbH is incorporated in Germany. Company registration number: HRB
54723, Amtsgericht Hamburg.
Registered Office: Wiesenkamp 22b, 22359 Hamburg, Germany.
Geschäftsführer: Jochen Witte.
Re: broken tests in TCK testsuite
Posted by Craig L Russell <Cr...@Sun.COM>.
Hi Christian,
From 11.1 of the specification:
The PersistenceManagerFactory implementation class must implement a no-
args
constructor. This allows the standard Java 6 service discovery
implementation to create a
factory instance upon which the getPersistenceManagerFactory methods
can be in-
voked.
I don't recall when this text was added. I have the history if it's
important to you.
Best regards,
Craig
On Jul 30, 2009, at 4:27 AM, CRomberg@versant.com wrote:
> Hi all,
>
> I hope this is the right list to report broken tests in the TCK
> testsuite,
> I'm referring to TCK 2.0, ignore my post if these issues are resolved
> meanwhile:
>
> The following tests are broken, because they rely on a no-args ctor
> requirement:
>
> org.apache.jdo.tck.api.persistencemanagerfactory.SetConnectionPassword
> org.apache.jdo.tck.api.persistencemanagerfactory.SetConnectionURL
> org.apache.jdo.tck.api.persistencemanagerfactory.SetConnectionUserName
> org.apache.jdo.tck.api.persistencemanagerfactory.SetIgnoreCache
> org.apache.jdo.tck.api.persistencemanagerfactory.SetMultithreaded
> org
> .apache.jdo.tck.api.persistencemanagerfactory.SetNonTransactionalRead
> org
> .apache.jdo.tck.api.persistencemanagerfactory.SetNonTransactionalWrite
>
> org.apache.jdo.tck.api.persistencemanagerfactory.SetOptimistic
> org.apache.jdo.tck.api.persistencemanagerfactory.SetRetainValues
>
> Per specification the instantiation of the vendor implementation of
> javax.jdo.PersistenceManagerFactory is
> implementation specific, so Class.newInstance() is not valid.
> If there is any JDOHelper.getPersistenceManagerFactory() overload
> which
> returns a mutable PMF, the tests
> can be rewritten using this. Otherwise the tests can not be fixed.
>
> Best regards,
>
> Christian
>
>
>
> ---
> Versant GmbH
> European Headquarters
>
> Christian Romberg
> Wiesenkamp 22 b
> D 22359 Hamburg
> ---
> Versant GmbH is incorporated in Germany. Company registration
> number: HRB
> 54723, Amtsgericht Hamburg.
> Registered Office: Wiesenkamp 22b, 22359 Hamburg, Germany.
> Geschäftsführer: Jochen Witte.
Craig L Russell
Architect, Sun Java Enterprise System http://db.apache.org/jdo
408 276-5638 mailto:Craig.Russell@sun.com
P.S. A good JDO? O, Gasp!
Re: broken tests in TCK testsuite
Posted by CR...@versant.com.
Hi Andy,
I've checked 2.0, it's not in there, so it must have been introduced
post-2.0.
However I still have concerns with this approach:
Normally PMFs are constructed via the static helper method contained in
the vendor's PMF implementation.
This introduces an alternative code path, which e.g. in our implementation
is not the same, because
our BootstrapPMF implementation decides based on certain properties which
PMF implementation is actually
used.
Best regards,
Christian
From:
Andy Jefferson <an...@datanucleus.org>
To:
jdo-dev@db.apache.org
Cc:
CRomberg@versant.com, Christian Ernst <ce...@versant.com>
Date:
30.07.2009 14:39
Subject:
Re: broken tests in TCK testsuite
> I hope this is the right list to report broken tests in the TCK
testsuite,
> I'm referring to TCK 2.0, ignore my post if these issues are resolved
> meanwhile:
> The following tests are broken, because they rely on a no-args ctor
Hi Christian,
Maybe I'm missing something but if I look at spec (JDO2.2 in my case, but
not
sure what version it appeared in - likely post-JDO2.0) section 11.1
<spec>
The PersistenceManagerFactory implementation class must implement a
no-args
constructor. This allows the standard Java 6 service discovery
implementation
to create a factory instance ...
</spec>
--
Andy
DataNucleus (http://www.datanucleus.org)
Re: broken tests in TCK testsuite
Posted by Andy Jefferson <an...@datanucleus.org>.
> I hope this is the right list to report broken tests in the TCK testsuite,
> I'm referring to TCK 2.0, ignore my post if these issues are resolved
> meanwhile:
> The following tests are broken, because they rely on a no-args ctor
Hi Christian,
Maybe I'm missing something but if I look at spec (JDO2.2 in my case, but not
sure what version it appeared in - likely post-JDO2.0) section 11.1
<spec>
The PersistenceManagerFactory implementation class must implement a no-args
constructor. This allows the standard Java 6 service discovery implementation
to create a factory instance ...
</spec>
--
Andy
DataNucleus (http://www.datanucleus.org)