[
https://issues.apache.org/jira/browse/HADOOP-12048?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14608894#comment-14608894
]
Allen Wittenauer commented on HADOOP-12048:
-------------------------------------------
bq. What is the best way to identify how this failed?
FYI, we've fixed this newer versions of test-patch, but they aren't on hadoop's
trunk yet. You can always force Jenkins to switch to branch HADOOP-12111 and
triggering the newer version by doing two things:
a) naming your patch blah.HADOOP-12111.xx.patch
b) changing something in test-patch.sh as part of that patch (such as adding a
line)
be aware that test-patch.sh in HADOOP-12111 is clearly different, so make it a
simple, not-likely-to-conflict change like adding a line after the bang path
> many undeclared used dependencies (and declared unused dependencies)
> --------------------------------------------------------------------
>
> Key: HADOOP-12048
> URL: https://issues.apache.org/jira/browse/HADOOP-12048
> Project: Hadoop Common
> Issue Type: Bug
> Components: build
> Affects Versions: 3.0.0
> Reporter: Sangjin Lee
> Assignee: Gabor Liptak
> Attachments: HADOOP-12048.1.patch, HADOOP-12048.2.patch,
> HADOOP-12048.3.patch, dependency-analysis.txt, hadoop-unused.txt
>
>
> Currently there are numerous undeclared used dependencies and declared unused
> dependencies in the hadoop projects.
> Undeclared used dependencies are easier errors to correct, and correcting
> them will lead to a better management of dependencies (enabling stricter
> dependency frameworks down the road).
> Declared unused dependencies are harder to resolve, as many may be legitimate
> runtime dependencies. But fixing them would lead to smaller profiles for
> hadoop projects.
> We can do a one-time scan of dependency issues and fix them. However, in the
> long run, it would be nice to be able to enforce those rules via maven
> plug-in.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
