I agree, this is becoming unmanageable. Do we think that post HBase 1.1 and us being on supported APIs (even in coprocessor land) that we'll need a shim layer? If likely not, then our branches will shrink over time (hopefully sooner rather than later), so maybe we can live with the current strategy. Otherwise, I agree that it's time to rethink this.
On Wed, May 27, 2015 at 1:30 PM, Nick Dimiduk <ndimi...@apache.org> wrote: > We now have 7 (!!!) active branches. Is there a plan to deprecate any of > them in the near future -- will there be a 4.5-HBase-1.0 release, for > example? Perhaps it's worth re-evaluating the cost-benefit analysis of a > shim layer. > > -n