[
https://issues.apache.org/jira/browse/HADOOP-7554?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13087399#comment-13087399
]
Alejandro Abdelnur commented on HADOOP-7554:
--------------------------------------------
+100 to consistency.
common and coming HDFS are consistent already and I from what i read in between
lines Eric's intentions for mapreduce will ride on this consistency.
Using assembly:single is expected by Maven developers. Still, there is a little
problem with it, it is not bound to any lifecycle phase, thus it is not easy to
combine it with other tasks. For Hadoop common and hdfs (and I assume
mapreduce) we have to so some tricks before and after assembly:single to handle
symlinks. The easiest way is doing as we are doing today in common and hdfs.
The more complex way (for which will need a Maven plugin) is to create a new
lifcycle with multiple phases and use it to bind the assembly:single and the
pre/post tasks (not rocket science but requires some Maven crafting)
> Use 'assembly' rather than '-Ptar' for creating the tarball
> -----------------------------------------------------------
>
> Key: HADOOP-7554
> URL: https://issues.apache.org/jira/browse/HADOOP-7554
> Project: Hadoop Common
> Issue Type: Improvement
> Components: build
> Reporter: Arun C Murthy
>
> AFAIK assembly is the way to create artifacts - currently we use profiles
> (-Ptar or -Pbintar).
> MapReduce uses assembly via MAPREDUCE-279.
> I'd like to see a common way across the whole project. Thoughts?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
