Hi Jacques, all,

Currently, we can't avoid having duplicates in test data when we load such
data just before the execution a test-suite/test-case. We should not
concern ourselves to much with this. After all it is just data for test,
and reloading a few duplicates should not be regarded as a major issue.

However, if the community is adamantly set on removing such duplicates,
then it should work on having test-data being loaded before any and all
test-suites/test-cases gets executed. IMO this involves moving test-data
from within the testdef folder (like in the order component) to the data
folder of the component and having a separate loadTestData task.

Best regards,

Pierre Smits

*Apache Trafodion <https://trafodion.apache.org>, Vice President*
*Apache Directory <https://directory.apache.org>, PMC Member*
Apache Incubator <https://incubator.apache.org>, committer
*Apache OFBiz <https://ofbiz.apache.org>, contributor (without privileges)
since 2008*
Apache Steve <https://steve.apache.org>, committer


On Sat, Apr 27, 2019 at 3:38 PM Jacques Le Roux <
[email protected]> wrote:

> Thanks Suraj,
>
> Can't we avoid the duplicated data?
>
> Jacques
>
> Le 27/04/2019 à 15:17, Suraj Khurana a écrit :
> > Hello team,
> >
> > I have checked and found that there is a data dependency of
> > workEffortId=9000 in the test case which is available in
> plugins/projectmgr
> > component.
> >
> > This was the main reason testIntegration was failing without having
> plugins
> > component. I will take care of it and add respective dependent data on
> > order test data file.
> >
> > I think its making sense now and we don't need to revert now.
> >
> > --
> > Best Regards,
> > Suraj Khurana
> >
> >
> >
> >
> >
> >
> > On Sat, Apr 27, 2019 at 10:14 AM Suraj Khurana <[email protected]>
> > wrote:
> >
> >> Sure Jacques,
> >>
> >> I am into it today and if got nothing I will remove OrderTests.groovy
> >>
> >> --
> >> Best Regards,
> >> Suraj Khurana
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >> On Fri, Apr 26, 2019 at 7:27 PM Jacques Le Roux <
> >> [email protected]> wrote:
> >>
> >>> Hi Suraj,
> >>>
> >>> I think that, as suggested by Mathieu, in the meantime it's better to
> >>> remove “OrderTests.groovy”
> >>>
> >>> Because it could hide other issues else reported by Buildbot which is
> our
> >>> last safeguard
> >>>
> >>> Thanks
> >>>
> >>> Jacques
> >>>
> >>> Le 25/04/2019 à 10:52, Pierre Smits a écrit :
> >>>> Hi Mathieu,
> >>>>
> >>>> Is there a way to move this forward?
> >>>>
> >>>> Best regards,
> >>>>
> >>>> Pierre Smits
> >>>>
> >>>> *Apache Trafodion <https://trafodion.apache.org>, Vice President*
> >>>> *Apache Directory <https://directory.apache.org>, PMC Member*
> >>>> Apache Incubator <https://incubator.apache.org>, committer
> >>>> *Apache OFBiz <https://ofbiz.apache.org>, contributor (without
> >>> privileges)
> >>>> since 2008*
> >>>> Apache Steve <https://steve.apache.org>, committer
> >>>>
> >>>>
> >>>> On Sat, Apr 20, 2019 at 2:25 PM Pierre Smits <[email protected]>
> >>> wrote:
> >>>>> Maybe we should move the load aspects regarding tests out of the test
> >>>>> suite invocations altogether.
> >>>>> The gradlew tasks states:
> >>>>>
> >>>>> task testIntegration(group: ofbizServer) {
> >>>>>
> >>>>> dependsOn 'ofbiz --test'
> >>>>>
> >>>>> description 'Run OFBiz integration tests; You must run loadAll before
> >>>>> running this task'
> >>>>>
> >>>>> }
> >>>>>
> >>>>>
> >>>>> IMO, loading test data could be part of the loadAll task.
> >>>>>
> >>>>>
> >>>>> Best regards,
> >>>>>
> >>>>> Pierre Smits
> >>>>>
> >>>>> *Apache Trafodion <https://trafodion.apache.org>, Vice President*
> >>>>> *Apache Directory <https://directory.apache.org>, PMC Member*
> >>>>> Apache Incubator <https://incubator.apache.org>, committer
> >>>>> *Apache OFBiz <https://ofbiz.apache.org>, contributor (without
> >>> privileges)
> >>>>> since 2008*
> >>>>> Apache Steve <https://steve.apache.org>, committer
> >>>>>
> >>>>>
> >>>>> On Sat, Apr 20, 2019 at 1:56 PM Mathieu Lirzin <
> >>> [email protected]>
> >>>>> wrote:
> >>>>>
> >>>>>> Pierre Smits <[email protected]> writes:
> >>>>>>
> >>>>>>> I believe there are a few more where testing individual test-suites
> >>>>>> and/or
> >>>>>>> test-cases are dependent on data loaded in other test-suites and/or
> >>>>>> other
> >>>>>>> test-cases.
> >>>>>> I have the same experience.  Moreover another source of fragility is
> >>>>>> that tests depend on other tests within a single OFBiz “test-case”,
> >>>>>> meaning one test can depend on the data produced by another test.
> >>> This
> >>>>>> is acceptable for a “simple-method-test” because the order of
> >>> execution
> >>>>>> is sequential and managed by OFBiz, but this is problematic for
> JUnit
> >>>>>> tests (Groovy, Java) because the order while being deterministic
> >>> depends
> >>>>>> on the arbitrary order imposed by the JVM.
> >>>>>>
> >>>>>> For example I know for a fact that “QuoteTests.groovy” is suffering
> >>> from
> >>>>>> that issue.
> >>>>>>
> >>>>>>> While I don't hear/read about failing testIntegration (except where
> >>>>>> code in
> >>>>>>> the base is faulty, not when test-suites/cases are faulty), I see
> >>>>>> following
> >>>>>>> failures in test executions in OFBiz against jdk11:
> >>>>>>>
> >>>>>>>
> >>>>>>>      1. Execution failed for task ':ofbiz --test component=webapp
> >>> --test
> >>>>>>>      suitename=webapptests'.
> >>>>>>>      2. Execution failed for task ':ofbiz --test
> component=accounting
> >>>>>> --test
> >>>>>>>      suitename=invoicetest'.
> >>>>>>>      3. Execution failed for task ':ofbiz --test component=order
> >>> --test
> >>>>>>>      suitename=ordertests'.
> >>>>>>>      4. Execution failed for task ':ofbiz --test component=product
> >>> --test
> >>>>>>>      suitename=producttests'.
> >>>>>>>
> >>>>>>> Do we have these test failing also when doing the test execution
> >>> against
> >>>>>>> jdk8?
> >>>>>>> *Caveat: I recently set this up, so there may still be some
> >>>>>> configuration
> >>>>>>> issues in the jdk11-test setup.. *
> >>>>>> I have just tested the “ordertests” test-suite with Icedtea 3.7
> >>> (jdk-8)
> >>>>>> and it is still failing, so it seems unrelated in that case.
> >>>>>>
> >>>>>> Thanks.
> >>>>>>
> >>>>>> --
> >>>>>> Mathieu Lirzin
> >>>>>> GPG: F2A3 8D7E EB2B 6640 5761  070D 0ADE E100 9460 4D37
> >>>>>>
>

Reply via email to