On Sep 2, 2009, at 10:27 AM, Abhijit Kumar wrote:
> Lloyd does not work on wednesdays. I will call him and see if he can
> look at it.
OK.
>
> Did the quicklook tests pass in the development workspace before the
> commit?
We didn't run the QLs: normally we just run the ORB integration
tests. We should probably add the QLs to
our ORB work.
> According to hudson logs, the tests were passing on the trunk before
> the commit for corba integration.
Yes, that's true. But I can't find any evidence in the log that the
ORB commit had anything to do with the
failures. I'm not sure the ORB is even running for those tests.
>
> Was gmbal updated as part of this commit?
There are no recent gmbal commits. The last commit was on 8/24 which
was for gmbal version 3.0.0-b011,
which was also pushed to maven on 8/24.
> AMX depends on gmbal, maybe that caused the test failures.
Actually AMX depends on management-api, not on gmbal. gmbal depends
on management-api as well.
(We need to be careful here about the gmbal project on Kenai, which
contains both the gmbal bundle
and the management-api bundle).
management-api was last updated on 8/31 to version 3.0.0-b005 by
Prashant.
Ken.
>
> Ken Cavanaugh wrote:
>>
>>
>> On Sep 2, 2009, at 9:43 AM, Abhijit Kumar wrote:
>>
>>> Thanks Ken.
>>>
>>> There are some quicklook test failures after the corba
>>> integration. Can someone look at the failures and address them?
>>
>> So far I can't see any evidence that the ORB caused the problems,
>> or even that the ORB is running during these tests.
>> What I do see at http://hudson.glassfish.org/job/gf-trunk-build-continuous
>> for build #2185 is a bunch of admin-related
>> failures, including a lot of InstanceNotFoundExceptions for
>> ObjectName amx:pp=,type=domain-root. That seems
>> very strange, as that is the ObjectName of the amx root that should
>> be created very early in GF startup.
>>
>> Lloyd, could you take a look at this?
>>
>> Also, the latest build (#2188) failed completely, with cannot find
>> symbol errors for HttpParserState. I know that is not
>> ORB related!
>>
>> The ORB may have contributed somehow, but there are definitely
>> other issues here that need to be investigated.
>>
>> Thanks,
>>
>> Ken.
>>
>>>
>>> Ken Cavanaugh wrote:
>>>>
>>>> Darani just finished committing the required changes for the
>>>> latest ORB into GFv3. The changes
>>>> went into revision 31085, and include:
>>>> Updating glassfish.corba.version to 3.0.0-b024 in pom.xml
>>>> Updating security/ejb.security/src/main/java/com/sun/enterprise/
>>>> iiop/security/CSIV2TaggedComponentInfo.java to use CDRInputObject/
>>>> CDROutputObject instead of CDRInputStream/CDROutpuStream (which
>>>> no longer exist in the ORB)
>>>> Updating orb/orb-iiop/src/main/java/org/glassfish/enterprise/iiop/
>>>> impl/IIOPSSLSocketFactory to use CorbaAcceptor instead of
>>>> Acceptor (which also no longer exists in the ORB)
>>>> Updating orb/orb-iiop/src/main/java/org/glassfish/enterprise/iiop/
>>>> util/S1ASThreadPoolManager.java to avoid the
>>>> DEFAULT_INACTIVITY_TIMEOUT constant, which is not currently
>>>> publicly available
>>>> I'll probably move this once we do the changes to make the ORB
>>>> threadpool available in the glassfish-corba-orbgeneric bunde.
>>>> The ORB integration tests and all of the ORB unit tests have
>>>> passed for this, so we are done for SCF.
>>>> Obviously there is still some work left for HCF (hopefully much
>>>> sooner than that HCF date):
>>>> We don't have 8474 fixed until we understand why the ORB cannot
>>>> obtain the PackageAdmin interface (Darani will run some tests on
>>>> this tomorrow)
>>>> I've started working on the orb-iiop changes to integrate the ORB
>>>> support for lazy init, but that cannot make it in tonight. At
>>>> this point, the ORB changes to support lazy init have been
>>>> tested, so we should be close on this
>>>> I need to move the threadpool code into glassfish-corba-orb, and
>>>> then modify some other bundles in GFv3 to work with the new
>>>> location
>>>> I need to work on the source build support
>>>> Ken.
>>>
>>
>