On 07/19/2010 03:26 PM, Jeff Hammerbacher wrote:
AVRO-230 (shared directory with test schemas) would be great, but I'm not sure it's a blocker. Do we want to make AVRO-570 (Python implementation of MapReduce) a blocker? I can see the pros (suss out sharp edges in the implementation) and cons (it's a new feature that's going to take a bit of work to implement).
As bad as I want this, in general I don't like to hold off releases for a feature if there are already some other useful new features that are ready. However if waiting a week or two will make a difference then we should consider it. Has someone committed to working on AVRO-570?
FYI, I've talked to Flip from InfoChimps, and he's hoping to start work on a Ruby implementation of MapReduce sometime soon. It'd be wonderful to have either of these in 1.4.0, but if they don't make that, we can always make another release soon after that does include them.
I've added AVRO-588 (Avro usable and testable on Python 2.4). I'll need to augment the Python RPC to handle system errors correctly after AVRO-573 goes in, so I've filed AVRO-599 as a blocker as well. Lastly, I've been meaning to get to AVRO-528 (Twisted support) and AVRO-283 (change Python implementation from dictionaries to objects) forever, so the release seems as good as ever.
Would you like me to try to knock off AVRO-573 this week? Doug
