[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
RE: [LIKELY JUNK]RE: [eclipselink-dev] Running EclipseLink server tests without weaving?
|
Hi Tom,
> In the absence of dynamic weaving, I believe the most-likely case
> for users
> of production-level applications will be to run with static weaving
> (rather than
> now weaving). I suggest the tests in that configuration.
Yes, that would have been my second step.
> There are some already some targets in our ant build that do the static
> weaving, so it should be possible to leverage them.
Thanks for the hint, Tom. I will have a look!
Best regards,
Sabine
--
Sabine Heider
SAP AG
Pflichtangaben/Mandatory Disclosure Statements:
http://www.sap.com/company/legal/impressum.epx
> -----Original Message-----
> From: eclipselink-dev-bounces@xxxxxxxxxxx [mailto:eclipselink-dev-
> bounces@xxxxxxxxxxx] On Behalf Of Tom Ware
> Sent: Mittwoch, 18. November 2009 19:51
> To: Dev mailing list for Eclipse Persistence Services
> Subject: Re: [LIKELY JUNK]RE: [eclipselink-dev] Running EclipseLink
> server tests without weaving?
>
> Hi Sabine,
>
> In the absence of dynamic weaving, I believe the most-likely case
> for users
> of production-level applications will be to run with static weaving
> (rather than
> now weaving). I suggest the tests in that configuration. There are
> some
> already some targets in our ant build that do the static weaving, so it
> should
> be possible to leverage them.
>
> The above could be combined with the suggestion you have made about
> making
> the weaving flag explicit.
>
> -Tom
>
> Heider, Sabine wrote:
> > Hi James,
> >
> >
> >
> > thanks for your reply, that explains my observation.
> >
> >
> >
> > Unfortunately, the NetWeaver server doesn't support dynamic weaving
> > using class transformers either. We have already addressed this issue
> > some time ago without success. I will file yet another bug for
> NetWeaver
> > explaining that EclipseLink's weaving relies on that but honestly I
> > don't expect any positive response unless we have a "real" business
> > case... So let's work with the assumption that NetWeaver doesn't
> support
> > dynamic weaving at all in the near future.
> >
> >
> >
> > I don't think that there is a need to disable weaving automatically
> > inside the NetWeaver server. There is already a mechanism to disable
> > weaving explicitly in the persistence.xml file, so we can document
> that
> > the property "eclipselink.weaving" has to be set to "false" or
> "static"
> > for the NetWeaver platform. (It would be cool if the server platform
> got
> > a chance to check this actively and issue a warning if not set, but
> that
> > would be only a nice-to-have feature.) I personally prefer explicit
> > declaration to any kind of magic here.
> >
> >
> >
> > That would mean, however, that the server tests need explicit
> > declaration of the weaving behavior as well. I would propose the
> > following approach (please let me know if you think it is feasible):
> >
> > - All persistence.xml files that are used for the server
> tests
> > get a property <property name="eclipselink.weaving"
> > value="%%server-weaving%%"/>
> >
> > - build.properties gets a property server.weaving=true;
> > netweaver.properties (or any other server not supporting dynamic
> > weaving) overwrites this with server.weaving=false
> >
> > - build.xml replaces the token %%server-weaving%% with the
> > value of the property server.weaving
> >
> > What do you think?
> >
> >
> >
> > If no one objects, I would try it out and provide a patch in case it
> works.
> >
> >
> >
> > Best regards,
> >
> > Sabine
> >
> >
> >
> >
> >
> > *From:* eclipselink-dev-bounces@xxxxxxxxxxx
> > [mailto:eclipselink-dev-bounces@xxxxxxxxxxx] *On Behalf Of *James
> Sutherland
> > *Sent:* Montag, 16. November 2009 16:29
> > *To:* Dev mailing list for Eclipse Persistence Services
> > *Subject:* [LIKELY JUNK]RE: [eclipselink-dev] Running EclipseLink
> server
> > tests without weaving?
> >
> >
> >
> > EclipseLink no longer has a dependency on a temp classloader being
> > provided. But as it seems weaving is still not working, it seems
> > Netweaver is not allowing for weaving in some other way?
> >
> >
> >
> > My guess is NetWeaver is either,
> >
> > - loading the entity classes before letting EclipseLink provide a
> class
> > transformer,
> >
> > - or, ignore the transformer that EclipseLink returns.
> >
> >
> >
> > Could you determine which is the case?
> >
> >
> >
> > It looks like we need to continue to disable weaving if there is no
> temp
> > class loader, as it seems to be an indication of general lack of
> server
> > support for performing any transformation of the entity classes.
> >
> >
> >
> > Could you also ensure a bug is logged on NetWeaver that it is not
> > compliant with the EJB 3 container requirements.
> >
> >
> >
> >
> >
> >
> >
> > -----Original Message-----
> > *From:* Heider, Sabine [mailto:sabine.heider@xxxxxxx]
> > *Sent:* Monday, November 16, 2009 10:17 AM
> > *To:* eclipselink-dev@xxxxxxxxxxx
> > *Subject:* [eclipselink-dev] Running EclipseLink server tests without
> > weaving?
> >
> >
> >
> > Hi,
> >
> >
> >
> > I'm trying to run the EclipseLink server tests inside a NetWeaver
> engine
> > using the NetWeaver platform from the incubator. As a result, I get
> lots
> > of errors that seem to be related with weaving, e.g.:
> >
> >
> >
> > javax.persistence.PersistenceException: Exception [EclipseLink-0]
> > (Eclipse Persistence Services - 2.0.0.qualifier):
> > org.eclipse.persistence.exceptions.IntegrityException
> > Descriptor Exceptions:
> > ---------------------------------------------------------
> >
> > Exception [EclipseLink-60] (Eclipse Persistence Services -
> > 2.0.0.qualifier):
> org.eclipse.persistence.exceptions.DescriptorException
> > Exception Description: The method [_persistence_setconcurrencyC_vh]
> or
> > [_persistence_getconcurrencyC_vh] is not defined in the object
> > [org.eclipse.persistence.testing.models.jpa.advanced.ConcurrencyA].
> > Internal Exception: java.lang.NoSuchMethodException:
> >
> org.eclipse.persistence.testing.models.jpa.advanced.ConcurrencyA._persi
> stence_getconcurrencyC_vh()
> > Mapping:
> org.eclipse.persistence.mappings.OneToOneMapping[concurrencyC]
> > Descriptor:
> >
> RelationalDescriptor(org.eclipse.persistence.testing.models.jpa.advance
> d.ConcurrencyA
> > --> [DatabaseTable(CONCURRENCYA)])
> >
> >
> >
> > The NetWeaver server is not able to provide a temporary classloader,
> so
> > dynamic weaving doesn't work. The NetWeaver platform contains the
> > following method (AFAIK taken from the JBoss platform), which should
> > indicate to EL that there is no temp classloader, but apparently that
> > mechanism doesn't work:
> >
> >
> >
> > public JPAClassLoaderHolder
> > getNewTempClassLoader(PersistenceUnitInfo puInfo) {
> >
> > ClassLoader realClassLoader = puInfo.getClassLoader();
> >
> > AbstractSessionLog.getLog().log(AbstractSessionLog.WARNING,
> > "persistence_unit_processor_sap_temp_classloader_bypassed",//
> >
> > puInfo.getPersistenceUnitName(), realClassLoader);
> >
> > return new JPAClassLoaderHolder(realClassLoader,
> > NO_TEMP_CLASS_LOADER);
> >
> > }
> >
> >
> >
> > So what is the correct mechanism to run the server tests with weaving
> > being disabled?
> >
> > I experimented with the settings in test.properties (test.agent=-
> Dignore
> > and test.weaving=-DTEST_NO_WEAVING=true), which had some effect on
> the
> > number of failing tests, but the NoSuchMethodExceptions remained. For
> > testing purposes, I edited the persistence.xml file of one of the
> test
> > applications to disable weaving, but that can't be the proper way to
> do
> > it...
> >
> > What would you suggest?
> >
> >
> >
> > Thanks and best regards,
> >
> > Sabine
> >
> >
> >
> >
> >
> > *Sabine Heider
> > **SAP AG*
> >
> > Pflichtangaben/Mandatory Disclosure Statements:
> > http://www.sap.com/company/legal/impressum.epx
> >
> >
> >
> >
> >
> >
> > ---------------------------------------------------------------------
> ---
> >
> > _______________________________________________
> > eclipselink-dev mailing list
> > eclipselink-dev@xxxxxxxxxxx
> > https://dev.eclipse.org/mailman/listinfo/eclipselink-dev
> _______________________________________________
> eclipselink-dev mailing list
> eclipselink-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/eclipselink-dev
