[
https://issues.apache.org/jira/browse/HADOOP-6671?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12853146#action_12853146
]
Lars Francke commented on HADOOP-6671:
--------------------------------------
I don't have a lot of insight into the Hadoop Common project itself but I've
done a lot of the work on the recent HBase transition together with Paul Smith
and Kay Kay and would like to offer my help if needed.
There are basically two ways to go: Either shuffle around a lot of directories
to conform to the standard Maven layout or to override the Maven defaults and
keep the current layout. Both ways are actually very painful but I prefer the
first one. While that means a lot of work and a lot of swearing it also means
that you'll get a consistent layout (with most other Maven projects) and the
configuration is a lot easier. Some tools tend(ed) to not work properly with
the non standard layout (strictly speaking this are bugs in the tools/plugins).
You basically have to do a lot of work up front but after that it shouldn't be
too hard to maintain.
The directory moving/renaming unfortunately tends not to work too well with
Subversion and branches (or I was doing it wrong) so I don't know how big the
benefit would be to start a new one for doing this.
What Paul Smith has done in HBASE-2099 is to provide a script that contained
all the necessary commands (svn mv...) to finish the move (there are a couple
more changes in other tickets) and a patch containing the .pom files. This has
been much improved since and we've learned a lot from it. We are in fact still
tweaking it and will change the pom structure once again to rely on the common
Apache parent pom in addition to a few smaller fixes that are still outstanding.
I'd propose a similar structure as HBase now has: A pom in the main directory
with common information (packaging "pom") and then two modules ("core" and
"contrib"). So the src/contrib directory would move to the top-level as would
the docs directory (btw: I've never used Forrest with Maven but if Ant can do
it Maven should be able to do it too). The final tarball as currently (trunk)
created by ant tar should be easily reproducible by Maven (I couldn't find
where most of the conribs ended up, are those currently missung from the .tar?)
Paul Smith has indicated interest in this and I'd be interested and willing to
help too if needed. But if you've already got someone to do the work I'm not
going to complain :) It'd just be a shame to duplicate work here as the moving
around of stuff in SVN is painful work if that's the way you decide to go.
> To use maven for hadoop common builds
> -------------------------------------
>
> Key: HADOOP-6671
> URL: https://issues.apache.org/jira/browse/HADOOP-6671
> Project: Hadoop Common
> Issue Type: Improvement
> Components: build
> Affects Versions: 0.22.0
> Reporter: Giridharan Kesavan
>
> We are now able to publish hadoop artifacts to the maven repo successfully [
> Hadoop-6382]
> Drawbacks with the current approach:
> * Use ivy for dependency management with ivy.xml
> * Use maven-ant-task for artifact publishing to the maven repository
> * pom files are not generated dynamically
> To address this I propose we use maven to build hadoop-common, which would
> help us to manage dependencies, publish artifacts and have one single xml
> file(POM) for dependency management and artifact publishing.
> I would like to have a branch created to work on mavenizing hadoop common.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
