Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hadoop >> mail # general >> Hadoop Mavenization - HADOOP-6671

Copy link to this message
Hadoop Mavenization - HADOOP-6671
After some work (much more than expected) and several iterations with folks
in the JIRA, HADOOP-6671 (https://issues.apache.org/jira/browse/HADOOP-6671)
should be ready for prime time. Thanks to Tom White helped getting Jenkins
and JIRA scripts working.

It would be great if folks do a test drive so it can be committed to trunk
and we would start the Mavenization of the rest of Hadoop (HADOOP-7412)

(Without wanting to start a religious discussion here) Using Maven for
building Hadoop brings several benefits:

* Testing is faster. Assuming TestConfiguration takes 0 secs, a warm run
using Ant takes 52 secs, a warm run using Maven takes 11 secs.
* Importing Hadoop Maven project from Eclipse/IntelliJ is trivial.
* Testcases run from the IDE without special settings.
* Publishing artifacts to the Maven repository becomes a simple command
(non-error prone task)
* The build system is driven by well defined steps, allowing developers to
understand/improve the build system.
* Versions of external dependencies for all Hadoop are maintained in one
place (hadoop-project/pom.xml)
* Building ALL Hadoop can be done in one step and all current code will be
* Most build tasks are faster using Maven (

The tests that use AOP are not integrated (I'll need some follow up help on
that in a follow up JIRA).

Said this, I guess we'll have some hiccups here and there as we all start
using this daily, as they show up I'll make sure we iron them out.