[
https://issues.apache.org/jira/browse/FLINK-1004?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14053525#comment-14053525
]
Stephan Ewen commented on FLINK-1004:
-------------------------------------
+1 good idea
Can all the repositories be added by default, such that maven initially
searches maven central, and goes though Clourera, MapR, Hortonworks to search
for dependencies not in Maven Central?
Could we also add build profiles for the different releases of those
distributions as well? Such that compiling the system for a specific vendor
distribution boils only down to setting the version.
If that does not work, we could enable repositories with a vendor property, and
then set the version with a different property.
Something like:
- {{-Dhadoop.vendor=<vendor>}}, where vendor can be "vanilla", "Cloudera",
"Hortonworks", "MapR". (although, isn't Hortonworks compatible with vanilla?)
- {{-Dhadoop.version=<version>}}
> Add MapR and Hortonworks repositories to main pom file.
> -------------------------------------------------------
>
> Key: FLINK-1004
> URL: https://issues.apache.org/jira/browse/FLINK-1004
> Project: Flink
> Issue Type: Improvement
> Components: Build System
> Reporter: Robert Metzger
> Assignee: Robert Metzger
> Priority: Minor
>
> As discussed here:
> http://apache-flink-incubator-mailing-list-archive.1008284.n3.nabble.com/How-to-compile-Flink-against-vendor-specific-Hadoop-libs-td977.html
--
This message was sent by Atlassian JIRA
(v6.2#6252)
