[ 
https://issues.apache.org/jira/browse/HADOOP-2690?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12563713#action_12563713
 ] 

Doug Cutting commented on HADOOP-2690:
--------------------------------------

The attached patch does not implement what I thought Enis suggested.  It 
creates a single jar with all the other jars flattened into it.  I'm not sure 
what the purpose of such a jar is and whether such a target belongs in Hadoop's 
build.xml.

I took Enis to be suggesting that we break the "compile-core-test" target in 
two, one with the 'javac' tasks and one with the 'jar' task.  There are a bunch 
of other tasks in "compile-core-test" that will need to be delt with too.  At a 
glance, I suspect most of these should remain in "compile-core-test" and just 
the 'jar' task should be moved to the new task, jar-core-test.  Then 
dependencies need to be adjusted so that things still work.


> Adding support into build.xml to build a special hadoop jar file that has the 
> MiniDFSCluster and MiniMRCluster classes among others necessary for building 
> and running the unit tests of Pig on the local mini cluster
> ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: HADOOP-2690
>                 URL: https://issues.apache.org/jira/browse/HADOOP-2690
>             Project: Hadoop Core
>          Issue Type: Improvement
>          Components: build
>    Affects Versions: 0.15.3
>            Reporter: Xu Zhang
>             Fix For: 0.17.0
>
>         Attachments: AddTestUtilsJarTarget.patch, build.xml, 
> hadoop-0.16.0-dev-test-utils.jar
>
>
> In order to build Pig and run its unit tests with ant, Pig needs to be built 
> and run against a hadoop jar file that not only has the MiniDFSCluster and 
> MiniMRCluster classes but also has classes from several 3rd party libraries.  
> I added a target to the build.xml file of Hadoop for that purpose (see the 
> "Compile code for Pig" section in the attached build.xml).   I also attached 
> the resulted hadoop jar with this report. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to