Perhaps branches similar to how hadoop 1.x and 2.x operate to handle the differences? Trunk can stay CDH3/1.x and a separate branch can be kept for hadoop 2.x until its more widely adopted?
On Thu, Oct 31, 2013 at 7:42 AM, Tim Williams <[email protected]> wrote: > On Sat, Oct 26, 2013 at 12:50 PM, Chris Rohr <[email protected]> wrote: > > It would be for running tests and possibly packaging (not sure if this is > > needed). It would not be needed for runtime. > > I dunno, it feels a heavy (separate install required) and random > precondition to building/packaging Blur to me. I guess I really like > that the quickstart now is simply 'git clone && mvn install'. What are > the alternatives? > > Thanks, > --tim >
