You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mrql.apache.org by "Leonidas Fegaras (JIRA)" <ji...@apache.org> on 2013/10/04 21:31:50 UTC

[jira] [Resolved] (MRQL-19) Fix license issues to prepare a new release candidate

     [ https://issues.apache.org/jira/browse/MRQL-19?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Leonidas Fegaras resolved MRQL-19.
----------------------------------

    Resolution: Fixed

MRQL-19-3.patch has been applied

> Fix license issues to prepare a new release candidate
> -----------------------------------------------------
>
>                 Key: MRQL-19
>                 URL: https://issues.apache.org/jira/browse/MRQL-19
>             Project: MRQL
>          Issue Type: Task
>    Affects Versions: 0.9.0
>            Reporter: Leonidas Fegaras
>            Assignee: Leonidas Fegaras
>             Fix For: 0.9.0
>
>         Attachments: MRQL-19.2.patch, MRQL-19-3.patch, MRQL-19.patch
>
>
> We need to address the license issues reported by sebb before we stage a new release candidate for vote.
> Here are sebb's emails:
> http://markmail.org/message/z66lnr5q4u24rwal
> http://markmail.org/message/k7zvajrn3hnla3ve
> My opinion is that we should not release any binary artifacts. We release only source. The problem is that the generated jars must bundle all the dependencies (which are the runtime JLine and CUP libraries) in order to run as single jars on Hadoop or Hama. This means that we have to repeat the JLine/CUP classes in the MRQL jars, which is something sebb doesn't like. Also, if we release binaries, the license files of the source distribution must be different from that of the binary since we don't use JLine/CUP source code.
> What do you think?
> All the other issues are easy to fix.
>  



--
This message was sent by Atlassian JIRA
(v6.1#6144)