Guys,

any reason there are no SNAPSHOT artifacts published for 0.94.1?
If this is a simple omission I can re-enable Apache Jenkins builds
to do so from this branch: http://svn.apache.org/viewvc/hbase/branches/0.94/

Please holler if you don't want it to happen (for whatever reason).

Thanks,
Roman.

On Wed, Jul 4, 2012 at 4:11 PM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:
> Guys,
>
> any reason there are no SNAPSHOT artifacts published for 0.94.1?
> If this is a simple omission I can re-enable Apache Jenkins builds
> to do so from this branch: http://svn.apache.org/viewvc/hbase/branches/0.94/
>
> Please holler if you don't want it to happen (for whatever reason).

Since I haven't heard much back I pushed the -SNAPSHOT artifacts
using the following Job:
   https://builds.apache.org/job/HBase-0.94-deploy/2/console

However, some of the new things in the 0.94.1 pom.xml seem
to make ivy-based projects unhappy. I see that Sqoop can't
be compiled because:
   * there's not default version of slf4j.version defined in the pom
   * junit.version == 4.10-HBASE-1

Do we really need that version of JUnit, I'd recommend trying
pretty hard to get rid of it given that the artifacts are not available
anywhere but in Gary Helmling's repo.

As for slf4j.version, can we just add a default value of 1.4.3
the same as with the default hadoop profile?

Thanks,
Roman.

P.S. Of course, it could be said that some of those Ivy projects
need to fix things on their end, but perhaps we can agree not
to break them unnecessarily?

The special junit version is for running tests in parallel.

Does any other Apache project run unit tests in parallel ?

Cheers

On Wed, Jul 4, 2012 at 10:30 PM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:

> On Wed, Jul 4, 2012 at 4:11 PM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:
> > Guys,
> >
> > any reason there are no SNAPSHOT artifacts published for 0.94.1?
> > If this is a simple omission I can re-enable Apache Jenkins builds
> > to do so from this branch:
> http://svn.apache.org/viewvc/hbase/branches/0.94/
> >
> > Please holler if you don't want it to happen (for whatever reason).
>
> Since I haven't heard much back I pushed the -SNAPSHOT artifacts
> using the following Job:
>    https://builds.apache.org/job/HBase-0.94-deploy/2/console
>
> However, some of the new things in the 0.94.1 pom.xml seem
> to make ivy-based projects unhappy. I see that Sqoop can't
> be compiled because:
>    * there's not default version of slf4j.version defined in the pom
>    * junit.version == 4.10-HBASE-1
>
> Do we really need that version of JUnit, I'd recommend trying
> pretty hard to get rid of it given that the artifacts are not available
> anywhere but in Gary Helmling's repo.
>
> As for slf4j.version, can we just add a default value of 1.4.3
> the same as with the default hadoop profile?
>
> Thanks,
> Roman.
>
> P.S. Of course, it could be said that some of those Ivy projects
> need to fix things on their end, but perhaps we can agree not
> to break them unnecessarily?
>

On Wed, Jul 4, 2012 at 11:04 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
> The special junit version is for running tests in parallel.

How is this different from the parallel runner available in JUnit 4.7+
    http://maven.apache.org/plugins/maven-surefire-plugin/examples/junit.html
?

> Does any other Apache project run unit tests in parallel ?

Sure. See above.

Thanks,
Roman.

P.S. The troublesome bit here is that HBase now depends on the
artifact that is not really reproducible. If there's a legitimate reason
of why migration to JUnit 4.7+ is not possible I think we have to
figure out how to depend on the required functionality without
this additional pain. Perhaps I can help with migration.

Hi Roman,

I monitor carefully the dependency we have on these private version:
see the backgroung in https://issues.apache.org/jira/browse/HBASE-4955
For both surefire and junit, we met bugs and fixed them. They're
committed in the junit trunk and surefire trunk.
For JUnit, we're waiting for the 4.11 release, the 4.10 being now 9
months old. Our fix was proposed 8 months ago, integrated to trunk 4
months ago.
We need it for the categories in surefire. Without this fix, JUnit
tries to fully initialize the classes when checking for categories.
Full story is here: https://github.com/KentBeck/junit/issues/359
In any case, HBase production code does not depend on them. This
dependency is only for the tests classes. If you don't run the hbase
tests, you can use the 4.10 version of JUnit.
If I understand well you want to push the artifact to a repo? Would it
make sense to use a specific maven profile? Such a profile already
exists actually. With "-P localTests", you desactivating categories.
It should be possible to set a JUnit 4.10 dependency on this existing
profile.
Cheers,,

N.

Roman,

I've tweaked that build so that the post build steps don't cause it look
like it failed.

Jon.

On Wed, Jul 4, 2012 at 10:30 PM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:

> On Wed, Jul 4, 2012 at 4:11 PM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:
> > Guys,
> >
> > any reason there are no SNAPSHOT artifacts published for 0.94.1?
> > If this is a simple omission I can re-enable Apache Jenkins builds
> > to do so from this branch:
> http://svn.apache.org/viewvc/hbase/branches/0.94/
> >
> > Please holler if you don't want it to happen (for whatever reason).
>
> Since I haven't heard much back I pushed the -SNAPSHOT artifacts
> using the following Job:
>    https://builds.apache.org/job/HBase-0.94-deploy/2/console
>
> However, some of the new things in the 0.94.1 pom.xml seem
> to make ivy-based projects unhappy. I see that Sqoop can't
> be compiled because:
>    * there's not default version of slf4j.version defined in the pom
>    * junit.version == 4.10-HBASE-1
>
> Do we really need that version of JUnit, I'd recommend trying
> pretty hard to get rid of it given that the artifacts are not available
> anywhere but in Gary Helmling's repo.
>
> As for slf4j.version, can we just add a default value of 1.4.3
> the same as with the default hadoop profile?
>
> Thanks,
> Roman.
>
> P.S. Of course, it could be said that some of those Ivy projects
> need to fix things on their end, but perhaps we can agree not
> to break them unnecessarily?
>

--
// Jonathan Hsieh (shay)
// Software Engineer, Cloudera
// [EMAIL PROTECTED]

+andrew (who can help us with more esoteric Maven needs)

Hi!

On Thu, Jul 5, 2012 at 1:31 AM, N Keywal <[EMAIL PROTECTED]> wrote:
> If I understand well you want to push the artifact to a repo? Would it
> make sense to use a specific maven profile? Such a profile already
> exists actually. With "-P localTests", you desactivating categories.
> It should be possible to set a JUnit 4.10 dependency on this existing
> profile.

Thanks for the very detailed explanation. I filed:
    https://issues.apache.org/jira/browse/HBASE-6332
and attached a patch with 2 simple workarounds.

Please let me know what do you think. The Bigtop's migration
to 0.94.1 is now blocked because of this:
   http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-Sqoop/102/

Thanks,
Roman.

I'm planning on committing Roman's workarounds soon, so please comment
if you have an objection or another solution.

On Thu, Jul 5, 2012 at 10:32 AM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:
> +andrew (who can help us with more esoteric Maven needs)
>
> Hi!
>
> On Thu, Jul 5, 2012 at 1:31 AM, N Keywal <[EMAIL PROTECTED]> wrote:
>> If I understand well you want to push the artifact to a repo? Would it
>> make sense to use a specific maven profile? Such a profile already
>> exists actually. With "-P localTests", you desactivating categories.
>> It should be possible to set a JUnit 4.10 dependency on this existing
>> profile.
>
> Thanks for the very detailed explanation. I filed:
>     https://issues.apache.org/jira/browse/HBASE-6332
> and attached a patch with 2 simple workarounds.
>
> Please let me know what do you think. The Bigtop's migration
> to 0.94.1 is now blocked because of this:
>    http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-Sqoop/102/
>
> Thanks,
> Roman.

--
Best regards,

   - Andy

Problems worthy of attack prove their worth by hitting back. - Piet
Hein (via Tom White)

I'm reading this from my phone, so I can't say, but I trust you guys ;-)
Le 5 juil. 2012 19:34, "Andrew Purtell" <[EMAIL PROTECTED]> a écrit :

> I'm planning on committing Roman's workarounds soon, so please comment
> if you have an objection or another solution.
>
> On Thu, Jul 5, 2012 at 10:32 AM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:
> > +andrew (who can help us with more esoteric Maven needs)
> >
> > Hi!
> >
> > On Thu, Jul 5, 2012 at 1:31 AM, N Keywal <[EMAIL PROTECTED]> wrote:
> >> If I understand well you want to push the artifact to a repo? Would it
> >> make sense to use a specific maven profile? Such a profile already
> >> exists actually. With "-P localTests", you desactivating categories.
> >> It should be possible to set a JUnit 4.10 dependency on this existing
> >> profile.
> >
> > Thanks for the very detailed explanation. I filed:
> >     https://issues.apache.org/jira/browse/HBASE-6332
> > and attached a patch with 2 simple workarounds.
> >
> > Please let me know what do you think. The Bigtop's migration
> > to 0.94.1 is now blocked because of this:
> >
> http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-Sqoop/102/
> >
> > Thanks,
> > Roman.
>
>
>
> --
> Best regards,
>
>    - Andy
>
> Problems worthy of attack prove their worth by hitting back. - Piet
> Hein (via Tom White)
>

Jesse +1ed it so that's good enough for me.  Committed.

    - Andy

On Thu, Jul 5, 2012 at 11:06 AM, N Keywal <[EMAIL PROTECTED]> wrote:
> I'm reading this from my phone, so I can't say, but I trust you guys ;-)
> Le 5 juil. 2012 19:34, "Andrew Purtell" <[EMAIL PROTECTED]> a écrit :
>
>> I'm planning on committing Roman's workarounds soon, so please comment
>> if you have an objection or another solution.
>>
>> On Thu, Jul 5, 2012 at 10:32 AM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:
>> > +andrew (who can help us with more esoteric Maven needs)
>> >
>> > Hi!
>> >
>> > On Thu, Jul 5, 2012 at 1:31 AM, N Keywal <[EMAIL PROTECTED]> wrote:
>> >> If I understand well you want to push the artifact to a repo? Would it
>> >> make sense to use a specific maven profile? Such a profile already
>> >> exists actually. With "-P localTests", you desactivating categories.
>> >> It should be possible to set a JUnit 4.10 dependency on this existing
>> >> profile.
>> >
>> > Thanks for the very detailed explanation. I filed:
>> >     https://issues.apache.org/jira/browse/HBASE-6332
>> > and attached a patch with 2 simple workarounds.
>> >
>> > Please let me know what do you think. The Bigtop's migration
>> > to 0.94.1 is now blocked because of this:
>> >
>> http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-Sqoop/102/
>> >
>> > Thanks,
>> > Roman.
>>
>>
>>
>> --
>> Best regards,
>>
>>    - Andy
>>
>> Problems worthy of attack prove their worth by hitting back. - Piet
>> Hein (via Tom White)
>>

--
Best regards,

   - Andy

Problems worthy of attack prove their worth by hitting back. - Piet
Hein (via Tom White)

On Thu, Jul 5, 2012 at 11:08 AM, Andrew Purtell <[EMAIL PROTECTED]> wrote:
> Jesse +1ed it so that's good enough for me.  Committed.

Thanks Andrew! I pushed the latest HBase snapshot artifacts and
Sqoop seems to be fine now.

Thanks,
Roman.