You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@batchee.apache.org by Scott Kurz <sk...@gmail.com> on 2014/08/27 13:35:24 UTC
Approach for testing core 'jbatch' function leveraging test utilities?
Hi, it's been interesting to spend some time looking around the BatchEE
project and look at all the integrations and extensions that you've been
adding !
I'm hoping to start contributing here by fixing some problems I know have
been inherited from the RI codebase.
So say we have a test that involves a sequence like a
start/stop/restart...and say it's designed to test something in the "core"
function in the 'jbatch' module.
The most obvious place based on existing tests would seem to me to put the
test itself in 'jbatch'. However, it might be nice to leverage the
batchee-test utilities.
But batchee-test already has a dependency on jbatch (looks like because it
uses the JAXB model).
Seems like we could either have a new module, "batchee-testcases" (or
something) leveraging batchee-test. Or another approach might be to pull
out the JAXB model into a separate module if that would allow jbatch to
depend on batchee-test.
What do you think?
Thanks,
Scott
Re: Approach for testing core 'jbatch' function leveraging test utilities?
Posted by Romain Manni-Bucau <rm...@gmail.com>.
if you mean jbatch/src/main/ it can be depending the amount of code.
Some part of batchee are not impl dependent (extensions and some
utility classes of batchee-test). Here duplication is generally
acceptable. If not - too much code - then we can make the part
explicitely dependent on batchee (for test module) but it should be
stated explicitely.
+1 for TCKs, that's more or less what we do everywhere (@apache I mean)
Romain Manni-Bucau
Twitter: @rmannibucau
Blog: http://rmannibucau.wordpress.com/
LinkedIn: http://fr.linkedin.com/in/rmannibucau
Github: https://github.com/rmannibucau
2014-08-28 12:36 GMT+02:00 Scott Kurz <sk...@gmail.com>:
> Romain, Mark,
>
> Thanks for the replies and the welcomes !
>
> Mark, regarding the TCK, we are planning to open the TCK for public
> contributions
> as part of the jbatch project...shortly. I hope to get around to this in
> the next few weeks.
>
> So if a bug happens to point to a notable TCK coverage gap, adding a TCK
> test would be
> an option and we'd be happy to have help contributing these tests.
>
> But I was still curious to hear where we'd add some additional non-TCK
> tests.
>
> Sounds like Romain is OK starting by duplicating some of the batchee-test
> classes .. in the core 'jbatch' module itself even?
>
> I'll next deliver a patch trying some approach for the next bug I had in
> mind from the RI:
> https://java.net/bugzilla/show_bug.cgi?id=5407
> (an NPE stopping not-yet-started partitions).
>
> Actually, regarding Mark's point, this could be a good TCK test candidate
> (or part of a set covering "incomplete" partitions).
>
> But also note it's a one-line null guard fix to the impl... so ideally it
> would be possible to commit fix + test without having to wait for a TCK
> update.
>
> Scott
>
>
>
>
>
>
>
>
> On Wed, Aug 27, 2014 at 8:10 AM, Mark Struberg <st...@yahoo.de> wrote:
>
>> Hi Scott!
>>
>> Good to have you around!
>>
>> Would that probably be a candidate for a JBatch TCK enhancement?
>> It would be great if all the effort also would lead to improving the TCK.
>>
>> If the answer is yes, then how can we contribute to the TCK in a smart way?
>> Quite a few of us are JCP members on other EGs, so that part should be
>> fine too.
>>
>> LieGrue,
>> strub
>>
>>
>>
>>
>>
>> > On Wednesday, 27 August 2014, 13:38, Scott Kurz <sk...@gmail.com>
>> wrote:
>> > > Hi, it's been interesting to spend some time looking around the BatchEE
>> > project and look at all the integrations and extensions that you've been
>> > adding !
>> >
>> > I'm hoping to start contributing here by fixing some problems I know have
>> > been inherited from the RI codebase.
>> >
>> > So say we have a test that involves a sequence like a
>> > start/stop/restart...and say it's designed to test something in the
>> > "core"
>> > function in the 'jbatch' module.
>> >
>> > The most obvious place based on existing tests would seem to me to put
>> the
>> > test itself in 'jbatch'. However, it might be nice to leverage the
>> > batchee-test utilities.
>> >
>> > But batchee-test already has a dependency on jbatch (looks like because
>> it
>> > uses the JAXB model).
>> >
>> > Seems like we could either have a new module, "batchee-testcases" (or
>> > something) leveraging batchee-test. Or another approach might be to
>> pull
>> > out the JAXB model into a separate module if that would allow jbatch to
>> > depend on batchee-test.
>> >
>> > What do you think?
>> >
>> > Thanks,
>> > Scott
>> >
>>
Re: Approach for testing core 'jbatch' function leveraging test utilities?
Posted by Scott Kurz <sk...@gmail.com>.
Romain, Mark,
Thanks for the replies and the welcomes !
Mark, regarding the TCK, we are planning to open the TCK for public
contributions
as part of the jbatch project...shortly. I hope to get around to this in
the next few weeks.
So if a bug happens to point to a notable TCK coverage gap, adding a TCK
test would be
an option and we'd be happy to have help contributing these tests.
But I was still curious to hear where we'd add some additional non-TCK
tests.
Sounds like Romain is OK starting by duplicating some of the batchee-test
classes .. in the core 'jbatch' module itself even?
I'll next deliver a patch trying some approach for the next bug I had in
mind from the RI:
https://java.net/bugzilla/show_bug.cgi?id=5407
(an NPE stopping not-yet-started partitions).
Actually, regarding Mark's point, this could be a good TCK test candidate
(or part of a set covering "incomplete" partitions).
But also note it's a one-line null guard fix to the impl... so ideally it
would be possible to commit fix + test without having to wait for a TCK
update.
Scott
On Wed, Aug 27, 2014 at 8:10 AM, Mark Struberg <st...@yahoo.de> wrote:
> Hi Scott!
>
> Good to have you around!
>
> Would that probably be a candidate for a JBatch TCK enhancement?
> It would be great if all the effort also would lead to improving the TCK.
>
> If the answer is yes, then how can we contribute to the TCK in a smart way?
> Quite a few of us are JCP members on other EGs, so that part should be
> fine too.
>
> LieGrue,
> strub
>
>
>
>
>
> > On Wednesday, 27 August 2014, 13:38, Scott Kurz <sk...@gmail.com>
> wrote:
> > > Hi, it's been interesting to spend some time looking around the BatchEE
> > project and look at all the integrations and extensions that you've been
> > adding !
> >
> > I'm hoping to start contributing here by fixing some problems I know have
> > been inherited from the RI codebase.
> >
> > So say we have a test that involves a sequence like a
> > start/stop/restart...and say it's designed to test something in the
> > "core"
> > function in the 'jbatch' module.
> >
> > The most obvious place based on existing tests would seem to me to put
> the
> > test itself in 'jbatch'. However, it might be nice to leverage the
> > batchee-test utilities.
> >
> > But batchee-test already has a dependency on jbatch (looks like because
> it
> > uses the JAXB model).
> >
> > Seems like we could either have a new module, "batchee-testcases" (or
> > something) leveraging batchee-test. Or another approach might be to
> pull
> > out the JAXB model into a separate module if that would allow jbatch to
> > depend on batchee-test.
> >
> > What do you think?
> >
> > Thanks,
> > Scott
> >
>
Re: Approach for testing core 'jbatch' function leveraging test utilities?
Posted by Mark Struberg <st...@yahoo.de>.
Hi Scott!
Good to have you around!
Would that probably be a candidate for a JBatch TCK enhancement?
It would be great if all the effort also would lead to improving the TCK.
If the answer is yes, then how can we contribute to the TCK in a smart way?
Quite a few of us are JCP members on other EGs, so that part should be fine too.
LieGrue,
strub
> On Wednesday, 27 August 2014, 13:38, Scott Kurz <sk...@gmail.com> wrote:
> > Hi, it's been interesting to spend some time looking around the BatchEE
> project and look at all the integrations and extensions that you've been
> adding !
>
> I'm hoping to start contributing here by fixing some problems I know have
> been inherited from the RI codebase.
>
> So say we have a test that involves a sequence like a
> start/stop/restart...and say it's designed to test something in the
> "core"
> function in the 'jbatch' module.
>
> The most obvious place based on existing tests would seem to me to put the
> test itself in 'jbatch'. However, it might be nice to leverage the
> batchee-test utilities.
>
> But batchee-test already has a dependency on jbatch (looks like because it
> uses the JAXB model).
>
> Seems like we could either have a new module, "batchee-testcases" (or
> something) leveraging batchee-test. Or another approach might be to pull
> out the JAXB model into a separate module if that would allow jbatch to
> depend on batchee-test.
>
> What do you think?
>
> Thanks,
> Scott
>
Re: Approach for testing core 'jbatch' function leveraging test utilities?
Posted by Romain Manni-Bucau <rm...@gmail.com>.
Hi Scott,
first let me say I'm very happy to see you here!
What I tried to do until today is to make the build "easy" (so i'm not
fan of pulling jaxb outside jbatch module) and make tests running
before deploying (speaking about maven phases).
It means - and AFAIK it is the case - we can have few duplication
dedicated to tests - think the waitForBatch utility is duplicated
IIRC.
IMHO jbatch module should really be seen as the kernel and can't
depends on the ecosystem around. I'm really not a fan of itests
modules since they build after the module you work in and that's the
best way to break something - if I'm alone please shout ;).
We can create test module from a shade of batchee-jbatch test jar if
the duplication becomes too important but I think it shouldn't be the
case.
Do you have some concrete example (pseudo code maybe) to try to think
more specifically on your case?
Romain Manni-Bucau
Twitter: @rmannibucau
Blog: http://rmannibucau.wordpress.com/
LinkedIn: http://fr.linkedin.com/in/rmannibucau
Github: https://github.com/rmannibucau
2014-08-27 13:35 GMT+02:00 Scott Kurz <sk...@gmail.com>:
> Hi, it's been interesting to spend some time looking around the BatchEE
> project and look at all the integrations and extensions that you've been
> adding !
>
> I'm hoping to start contributing here by fixing some problems I know have
> been inherited from the RI codebase.
>
> So say we have a test that involves a sequence like a
> start/stop/restart...and say it's designed to test something in the "core"
> function in the 'jbatch' module.
>
> The most obvious place based on existing tests would seem to me to put the
> test itself in 'jbatch'. However, it might be nice to leverage the
> batchee-test utilities.
>
> But batchee-test already has a dependency on jbatch (looks like because it
> uses the JAXB model).
>
> Seems like we could either have a new module, "batchee-testcases" (or
> something) leveraging batchee-test. Or another approach might be to pull
> out the JAXB model into a separate module if that would allow jbatch to
> depend on batchee-test.
>
> What do you think?
>
> Thanks,
> Scott