On Jan 9, 2015, at 1:10 PM, Vladimir Sitnikov <[email protected]> 
wrote:

>> I thought you mind just check in the MySQL dump as part of the script that 
>> creates the VM.
> 
> Do you mean check in all the dumps into calcite repository?
> I do not like this very much since it would increase the repository
> size and there is no easy way out.
> 
> I would rather prefer referencing some external blobs.
> Those blobs (dumps for mongodb, mysql, splunk, etc) are required just
> in case you test those adapters while the increase in repository size
> would hit all the developers.

How about this… put those resources into maven jars (similar to 
net.hydromatic.tpcds and pentaho.mondrian-data-foodmart-hsqldb and 
pentaho.mondrian-data-foodmart-json). You will need a small project to create 
each maven resource, and you can create a github repo for that. The raw data 
will of course be checked into the repo but very few people will ever need to 
pull or edit the git repo; they will use the maven artifact.

https://github.com/julianhyde/tpcds is an example of such a repo. 

If you create something similar we can “release” to conjars.org.

Julian

Reply via email to