sqoop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jarek Jarcec Cecho (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SQOOP-722) Improve Sqoop 2 packaging
Date Fri, 23 Nov 2012 18:26:59 GMT

    [ https://issues.apache.org/jira/browse/SQOOP-722?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13503261#comment-13503261

Jarek Jarcec Cecho commented on SQOOP-722:

Hi Abhijeet,
thank you very much for your feedback, I appreciate your time.

Regarding your concern on first suggestion. I believe that we quite did not understand each
other, so let me describe the suggestion little bit more. Right now user had to call maven
twice to build source and binary artifacts. With this patch he need to call maven only once.
However even thought we still build two separate artifacts. This suggestion is only about
building both artifacts during single maven call. Consider following list with created artifacts:

* mvn package -Dhadoop.profile=200
** sqoop-2.0.0-SNAPSHOT.tar.gz (source)
** sqoop-2.0.0-SNAPSHOT-bin-hadoop200.tar.gz (binary)
* mvn package -Dhadoop.profile=100
** sqoop-2.0.0-SNAPSHOT.tar.gz (source)
** sqoop-2.0.0-SNAPSHOT-bin-hadoop100.tar.gz (binary)

The source artifacts are exactly the same in both profiles as they are independent on Hadoop
version. They just happen to be created each time. Does this help clear your concern?

> Improve Sqoop 2 packaging
> -------------------------
>                 Key: SQOOP-722
>                 URL: https://issues.apache.org/jira/browse/SQOOP-722
>             Project: Sqoop
>          Issue Type: Improvement
>    Affects Versions: 2.0.0
>            Reporter: Jarek Jarcec Cecho
>            Assignee: Jarek Jarcec Cecho
>             Fix For: 2.0.0
>         Attachments: bugSQOOP-722.patch
> I would like to propose two improvements to current building (dist module):
> * Build both source and binary artifacts at the same time. As as I understand a lot of
other projects are doing that as well and it seems more friendly than creating one package
at the time.
> * Encode hadoop profile into binary artifact name. I believe that we should encode for
which Hadoop version is given binary artifact suitable. I know that only two jars are actually
"hadoop specific" whereas the others will work across all hadoop distributions, but I believe
that it will be more user friendly.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message