Hi Simon
have you tried to see if running T_TransSystem.java maybe increasing:
    static final int numReaderTasks         = ...
    static final int numWriterTasksA        = ...
    static final int numWriterTasksC        = ...
replicate the issue on your environment?

Also, I am running the tests on Linux 64-bits OS and Oracle JDK 1.6.x (64-bits)
(direct mode and memory mapped files).

Paolo


Simon Helsen wrote:
Thanks Paolo,

I am glad you agree.

I have currently nothing to reproduce issue 143 other than my own tests. I'll try to do something about this. As for 131, you can simply adopt the instrumentation I gave in the WI and see for yourself when running multiple queries/writes (at least 10 threads should do the job) that you run into issues.

Can you give that a try? Should be straightforward

Thanks

Simon


From:   Paolo Castagna <[email protected]>
To:     [email protected]
Date:   11/04/2011 12:40 PM
Subject:        Re: What are the JIRA issues to fix before a release?


------------------------------------------------------------------------



Hi Simon

Simon Helsen wrote:
 > Paolo,
 >
 > you cannot expect that bug reports always have a test case demonstrating
 > the problem.

Yes.

Not expecting anything, just asking politely: any help is appreciated.

 > That is not a reasonable expectation given that we are not
 > using Jena in a vanilla stand-alone manner, but integrate it a complex
 > piece of proprietary software.

Ack.

 > I provided quite a few detailest in 131 and
 > I would hope that if you take the defect serious, you investigate the
 > information I provided.

Yes.

None of the options I listed ignored JENA-131 or JENA-143.

I know it takes a lot of time and effort to produce a test to help
us in replicating the problem (and this is particularly true in case
of problems related to concurrency), but from the past bugs, once a
test case showing an issue is provided, bugs are fixed quite fast.

So, any help in making the issues described in JENA-131 and JENA-143
reproducible with some code we can run is really valuable and welcome.

Paolo

 >
 > thanks
 >
 > Simon
 >
 >
 >
 > From:
 > Paolo Castagna <[email protected]>
 > To:
 > [email protected]
 > Date:
 > 11/04/2011 10:17 AM
 > Subject:
 > Re: What are the JIRA issues to fix before a release?
 >
 >
 >
 > Simon Helsen wrote:
 >> I would like to see both
 >>
 >> https://issues.apache.org/jira/browse/JENA-131
 >> https://issues.apache.org/jira/browse/JENA-143
 >>
 >> resolved
 >>
 >> I provided input for 131 - I still have to try to produce a test-case
 > for
 >> 143, but depending on the origin of 131, it may be connected to the
 >> problem in 143
 >
 > Hi Simon
 >
 > One option in relation to JENA-131 and JENA-143, since they do not have
 > a test case we can use to reproduce the problem at the moment, we could
> decide to release TDB anyway and, as soon as the bug becomes reproducible,
 > we fix it and we can release a bug fix release x.y.(z+1).
 >
 > Another option is: we release all the Jena modules but TDB. For TDB we
 > produce a candidate release and wait a little bit more to push that out
 > to see if we can reproduce the problem (and fix it). However, in this
 > case Fuseki (and LARQ) will also be impacted.
 >
 > This way, neither JENA-131 or JENA-143 would be blockers for our first
 > Apache Jena release.
 >
 > I would not propose this normally and absolutely not if there were an
 > open bug with a test case to reliably reproduce it.
 >
 > Any help in writing a test case to reproduce the problem is certainly
 > appreciated. Thanks again for your feedback on JENA-131 and JENA-143.
 >
 > Paolo
 >
 >> thanks
 >>
 >> Simon
 >>
 >>
 >>
 >>
 >> From:
 >> Paolo Castagna <[email protected]>
 >> To:
 >> [email protected]
 >> Date:
 >> 11/04/2011 09:21 AM
 >> Subject:
 >> What are the JIRA issues to fix before a release?
 >>
 >>
 >>
 >> Hi,
 >> today, I have time to work on Jena.
 >>
 >> What are the JIRA issues to fix before a release?
 >>
 >> Paolo
 >>
 >>
 >>
 >>
 >
 >
 >
 >




Reply via email to