[
https://issues.apache.org/jira/browse/HADOOP-12556?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15169727#comment-15169727
]
Chris Nauroth commented on HADOOP-12556:
----------------------------------------
I'm slightly in favor of option 2: keep it as one big dir. That gives an easy
out if someone decides they really do want the whole world by putting
share/hadoop/tools/lib/* on the classpath. OTOH, I suppose we could come up
with a "whole world" shell profile that walks a more granular directory
structure and gathers everything.
In general, I really like the idea of using shell profiles to solve this
problem. We still have a gap in that we don't have equivalent functionality on
Windows. I have a hunch that it won't be feasible to offer all of the rich
features of the full shell rewrite in cmd, but maybe we can do just enough to
support classpath customization through profiles.
> KafkaSink jar files are created but not copied to target dist
> -------------------------------------------------------------
>
> Key: HADOOP-12556
> URL: https://issues.apache.org/jira/browse/HADOOP-12556
> Project: Hadoop Common
> Issue Type: Bug
> Reporter: Babak Behzad
> Assignee: Babak Behzad
> Attachments: HADOOP-12556.patch
>
>
> There is a hadoop-kafka artifact missing from hadoop-tools-dist's pom.xml
> which was causing the compiled Kafka jar files not to be copied to the target
> dist directory. The new patch adds this in order to complete this fix.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
