[
https://issues.apache.org/jira/browse/HCATALOG-76?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13080717#comment-13080717
]
Dmitriy V. Ryaboy commented on HCATALOG-76:
-------------------------------------------
+1 for having separate build targets for each set of drivers that pull in
non-trivial dependencies (HBase being a prime example).
Can we put contrib on Git or Google Code ("apache extras") to begin with, to
avoid the problems with contribs we ran into in Pig? Note that this would mean
we need to publish daily snapshots into a maven repo of some sort; I believe
there is some Apache infra support for that.
> Storage driver deployment
> -------------------------
>
> Key: HCATALOG-76
> URL: https://issues.apache.org/jira/browse/HCATALOG-76
> Project: HCatalog
> Issue Type: Improvement
> Reporter: Olga Natkovich
>
> We need to have a way to add new drivers to the system. This raises a few
> questions:
> (1) Where the drivers stored in SVN
> (2) What is the structure of deployed drivers (packages, directory structure,
> config, etc.)
> (3) How drivers get deployed
> We are working on the Hbase driver for HCat and starting to look at (1). Our
> proposal is to keep things simple and to create a driver directory in
> HCatalog source tree with a directory for each driver. Each driver would then
> have its own build process, dependencies etc. At the end of the build process
> it will produce whatever we decide is required for (2) but at least one or
> more jar files. We think all core drivers including the current default ones
> should follow the same structure (though we do not have to jump right away
> and re-factor the code). The core drivers would be the ones that the project
> agrees to maintain long term. We can also have similar structure under
> contrib for those that not initial considered to be core ones. Comments?
> We have some suggestions for (2) and (3) as well but want to discuss (1) first
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
